Management and phenotypic features of indigenous cattle in Rwanda

Breed improvement and conservation are optimally achieved when the available genetic resources are characterised and strategies developed to achieve the goals. This study aimed at investigating the management practices, performance and morphological features of the indigenous cattle ecotypes in Rwanda on 250 cattle farming households. A total of 20 measurements taken on 305 female and 45 male cattle were: body length (BL), height at withers (HW), leg height (LH), heart girth (HG), body weight (BW), tail length (TL), dewlap length (DL), dewlap width (DW), rump width (RW), ear length (EL), muzzle circumference (MC), horn length (HL), distance between horns (HS), hump length (HuL), hump width (HuW), navel depth (ND), udder length (UL), udder depth (UD), teat length (TL), and body condition score (BCS). Morphometric data was analysed by ecotype for each sex and age category since there were non-significant differences in geographical location. Results show that Rwanda has five types of indigenous cattle namely: Inyambo, Inkuku, Inkoromaijo, Inkungu and Bashi. The livestock system mostly used was extensive and household income was mainly from livestock. For Inyambo cattle, the popular ecotype, age at sexual maturity was 27.44±1.04 months for males and 28.76±1.02 months for female cows. Age at first calving was 33.8±0.83 months whereas calving interval was 13.60±0.45 months. Lactation length was found to be 6.84±0.29 months. The mean daily milk was 3.58±0.19 litres and the pre weaning calf survivability was 90±6.5%. Positive and high correlations were observed between BW, HG, HW, HuL, BL and HL. Indigenous cattle population of Rwanda are not homogenous by their morphological features and other productive traits, and therefore conservation will have to target the different ecotypes and this should be done with direct engagement of their keepers.


INTRODUCTION
The contribution of livestock to livelihoods of human Communities all over the world has been acclaimed (Hoffmann, 2010;Mwai et al., 2015).Indigenous cattle in particular form the backbone of relevant and sustainable *Corresponding author.E-mail: chirwa@gmail.comTel: +250-784-824250.Fax: +250-784824250.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License livestock production in most rangeland of the world, and in Eastern Africa they have been associated with gods and deities.Fittingly, when compared with their exotic counterparts, they are better adapted to survive and tolerate harsh environments.Based on this peculiarity of indigenous cattle, the ministry of agriculture in Rwanda initiated a program to conserve and ensure equitable utilization of indigenous animal genetic resources.The current global genetic diversity should be maintained being that millions of people directly depend on them for their livelihood (Rege and Gibson, 2003;Mekuriaw and Kebede, 2015;Muchenje et al., 2016).In Rwanda, cattle contribute significantly to livelihoods through provision to farm families with milk, meat, manure and cash income.They also play a significant role in the social and cultural values of the communities that rear them.Cattle produce a total of 703,436 metric tonnes (MT) of milk and 81,000 MT of meat annually in that country (NISR, 2015).
Indigenous cattle in the family of Bos indicus derived breeds present enhanced hardiness and disease resistance, low nutritional requirements and higher capability of feed utilization.However, their main products remain meat, milk, hides, and manure.In Rwanda, indigenous cattle are divided into five groups, namely: Inyambo, Inkuku, Bashi, Inkungu and Inkoromoijo.These have unique evolutionary adaptation to harsh climates (Hansen, 2004;Shabtay, 2015;Dossa and Vanvanhossou, 2016) and various endemic diseases (Murray et al., 1984;Mattioli et al., 2000).One of these adaptations is their documented tolerance to trypanosomosis (Roberts and Gray, 1973); a parasitic disease due to infection with Trypanosoma sp.whose vector is the tsetse flies.Another is the ability of these cattle to use poor-quality forage and limited quantities of feed.Zebu cattle (Bos indicus), are the majority of cattle types in Africa.They are adapted to dry environmental conditions and high temperatures and are known to be more resistant to tick infestation compared to Bos Taurus cattle (Mattioli et al., 2000;MRI, 2012).African zebu cattle inhabit western and eastern parts of Africa.Their large body size and high production levels in tsetse-free areas have made them more appealing to producers.The southern Africa region is largely made of cattle that are taurine, Sanga and their crosses (Hanotte et al., 2000).
Although, indigenous cattle genetic diversity remains large, cattle populations or breeds continue to face extinction.Unfortunately, this extinction is not restricted to cattle or livestock, but to other forms of biodiversity to the extent that globally, at least three species are lost per hour to inter alia urbanisation, deforestation and climate change (Byron-Cox, 2016).According to Rege and Tawah (1999) and Mekuriaw and Kebede (2015), 22% of African cattle breeds became extinct in the last century and 32% of the indigenous African cattle breeds are currently in danger of extinction (Guaza and Mamah, 2016).Particularly, the Rwandan indigenous cattle genetic resources are in danger of disappearing more rapidly than predicted, following uncontrolled crossbreeding and breed replacements with exotic breeds, especially of dairy function.This is exacerbated by the general lack of selection-based breed improvement programs while paradoxically; the demand for livestock products is continually increasing.Climate change is putting new pressures on livestock production, while the livestock contribute to the shift, through greenhouse emissions (Houghton and Callander, 1992;Nardone et al., 2010).Moreover, indigenous livestock although adapted to the local environments, are poor milk and meat producers compared to the commercial breeds raised in the extensive system (Renaudeau et al., 2012;Tsegaye et al., 2014;Dossa and Vanvanhossou, 2016).
More than ever, indigenous livestock genetic resources which constitute the largest proportion of livestock in those countries are increasingly being eroded through poorly planned crossbreeding and breed replacements and this ought to be arrested.Characterization of the indigenous cattle in Rwanda has generally been lacking, with just a few isolated studies having been documented hither to this study.One such study reported that there was molecular homogeneity among the different cattle populations in Rwanda (Ndumu et al., 2008) and this could be attributed to similarity in selection criteria identified earlier (Wurzinger et al., 2006).The gain for the country in conducting such studies is widespread public awareness of the resource that must not be lost.Rwandan farmers will gain by promoting good breeding practices for their indigenous animals, which can in the process promote agro-tourism.There is much more potential that can be exploited from these breeds but this can only happen if the animals are not lost.Through this study, we aimed to determine the phenotypic and morphometric characteristics of indigenous cattle breeds in Rwanda and decipher their distribution.Secondly, it was to assess their productive and reproductive performance under the current systems of production and management in order to enable cattle breeders and policy makers to make appropriate decision for their future utilization.

Study area and sampling procedure
The study used a descriptive, purposive and stratified survey design and was carried out in indigenous cattle keeping households of Rwanda from all the four provinces in such a way to capture the different zones of the country (Table 1.The Northern Province is highland, the eastern province is lowland, the southern province is midland while the western province is partly highland and partly lowland.Each province is made up of seven to ten districts.In each province, three districts were randomly selected and from each, three villages were selected from at least two sectors.The indigenous cattle characterized were located in highland (12%), midlands (44%) and lowland (44%).A ANGR-CIM tool was administered to indigenous cattle breeders to gather information on the indigenous cattle they rear.2).Live body weight and 20 body measurements were then taken on the selected animals.The studied animals were mature active breeders.Maturity was ascertained from possession of either three or four pairs of permanent teeth.Additionally, parity was also used for estimating the age of the cows while the number of years the bulls had been in use for breeding was also used for estimating age of bulls.Body morphometric measurements were done with the cattle placed in a confinement crush but while ensuring the animals were free standing as much as possible, to ensure accuracy of measurement.
A descriptor list of phenotypic characteristics to assist with the qualitative description of the animals, a colour chart to describe the coat colour of the animals, and calibrated tapes, a 50-cm spirit level (inclinometer), a pair of 50-cm callipers (outside diameter) and a measuring stick (hippometer) to measure the quantitative physical characteristics such as heart girth (HG) and body length (BL) were also used.The set of measurements taken for the traits considered in this study were as described by Brown et al. (1983), Adeyinka and Mohammed (2006), Kugonza et al. (2011a) and AU-IBAR (2015).

Data analysis
The filled questionnaires were then coded and entered into the SPSS (ver.16) computer software.Data analysis was thereafter Where y is the observation in breed i. μ is the overall mean, bi the effect of breed (i = 5), eij the random effect on the trait, independently and identically distributed with mean = 0 and variance = σ 2 .A total of 350 indigenous cattle were used for the characterization study across the country.Data was entered into SPSS software data capture system and cleaned of errors.Data analysis was then performed using Statistical Analysis Systems (SAS, version 9.2) software.The statistical model used for analysis was: Where, y is observation of the trait in farm type i, in altitude zone j, and for animal age k. µ = overall mean, fi = effect of farm type (i = 2), sj = agro-ecological system (j = 3), ak is the effect of age (l = 4), fj(sj) is the effect of interaction of farm type and agro-ecological system, eijkl is the random effect on the trait, independently and identically distributed with mean = 0 and variance =  2 .
Correlation coefficients between HG, HW, HuL, BL, HL, DW, LH, HuW, DL, MC, EL, BCS, HD, UL, RW, TeL, TaL, UD, ND and BW were computed on a within-sex basis to determine linear associations between them.Effort was made to ensure that the physiological status of cows sampled for the linear measurement analysis was similar so that the animals were evaluated as a homogenous group.Linear regression of HW on HG, HW on BL and HG on BW were conducted within each sex category of indigenous cattle.

Characteristics of indigenous cattle keepers' households and their herds
The majority of indigenous cattle rearing households were male-headed and only 16% were female-headed.The average number of cattle per household were 15.2±4.34 of Inyambo breed, 20.33±11.4Inkuku, 3.09±1.51Inkoromoijo, 0.48±0.14Bashi and 0.32±0.12Inkungu (Table 3).Household income was mainly earned from livestock (84%), crop (52%), business (28%) and salary (8%).Other than indigenous cattle, the households were also rearing chickens (12.5%), goats (8.4%), camels (4.2%) but not sheep (0.0%).There were six cattle management activities that the family members were involved in either singly or in various combinations.Sale of cattle was mostly performed by heads of households, followed by animal care particularly grazing, watering, and supplementary feeding from fodder banks (Table 4).
In effect, no family member solely implemented any one activity in isolation from other family members.The major farming activities are depicted in Table 4. Though all the farmers kept livestock, only 84% acknowledged livestock as the main source of income.Cattle farmers need to engage in other investments so as to earn enough to sustain their families.It is prudent that the farmers make a choice between purely indigenous cattle crossbreeds, or various levels of upgrades.This will enable them to be able to fit into specific production zones and systems (Mwacharo and Drucker, 2005;Kugonza et al., 2012b).Very few farmers were engaged in only one enterprise at the time of the study, with majority engaging in two or more enterprises.This lack of specialisation may be detrimental to targeting maximum productivity but continues to be a normal feature among smallholders who are more interested in sustainability of livelihoods mainly through diversification of animal and crop enterprises.

Grazing management and herd dynamics
The average land area for grazing of cattle for most farmers was five hectares or less (Table 5).This is land that is usually ideal for two high grade dairy cows (Mugasi et al., 1999).The major implication of this statistic is that farmers who have little land have to change priorities in terms of breed of choice and or even change the type of genotype they rear, and this change has to be immediate or in the near future.As showed by the type of grazing system, almost all the farmers keep their cattle under confinement, yet many of them still have access to open grasslands.The land use patterns will keep evolving as the land sizes also continuously change.
Majority of the farmers (68%) reported that they purchased the cows they own, and only 28% reported that they inherited and 4% acquired the cows as gifts (Table 5).This differs from findings in Uganda for the Ankole breed (Nabasirye et al., 2012) and for other breeds in Kenya (Rege et al., 2001).Whereas purchase is always associated with commercial-oriented nontraditional pastoralists, it was surprising in this study that inheritance and gifts that are strongly associated with traditional pastoralist communities make such a low contribution in Rwanda.The major shift could be attributed to the heavily commercialised livestock farming in Rwanda such that the socio-cultural practices of freely giving away of cattle have been eroded, and indeed, is the way to go for pastoralists to sustainably manage their stocks in the light of growing disease threat and growing economic demands for communities.Most farmers in the study (52%) reported not to carry out selective breeding of their cattle, yet they have some selection criteria for preferred traits.Interestingly, regarding acquisition, over one third of the cattle farmers acquired their first cattle when they were still children, and only 12.5% acquired cattle when 20 years or older (Table 5).All the interviewed respondents had had ancestors who reared local cattle, but what was poignant is that, a greater number reported great grandparents (34.8%) who reared these cattle, while fewer (19.6%) reported grand-parents and even fewer reported parents (8.7%) having reared indigenous cattle, and there was a drop by half in each generational change.
In general, cattle entered their current owners' herds mostly by birth or transfer (Table 5).Purchase of both cows and bulls, as well as receiving gifts of either cows or bulls are rare events among the farmers of indigenous cattle.This is rather surprising because these cattle still offer cultural roles in the communities where they are reared.Contemporary studies of similar cattle in other great lakes region countries (Ndumu et al., 2008;Kugonza et al., 2012b) show the opposite scenario, though in Kenya, low levels (6.6%) of use of cattle as gifts/dowry were found only for Masaai zebu cattle and absent for other breeds (Mwacharo and Drucker, 2005).On the other hand, exits from the herd are mainly through sale that is higher for male than female cattle.Death of cattle was reported by the farmers though the numbers were quite low, an indication of good health care.The other factor is the better tolerance of the indigenous breed(s) to the common diseases.It was very encouraging to find from the study that cattle thefts are almost absent for male cattle and completely absent for females, a rarity in the great lakes region where cattle rustling and thefts are widespread (Kugonza et al., 2012b).Credit for this should go to the strong law enforcement effort as well as the identification and animal traceability programmes that are widespread in Rwanda.
The number of cattle that exit the herd by being donated was higher than the number that is received for the same purpose.The mean exit per year per herd was 17.09 animals while the mean entry per year was 15.64 animals.The implication of this is that the national indigenous herd is shrinking since the direct removals/exits are more than the direct additions/entries.Removal is also indirectly a result of growing of crossbreeding.Therefore only direct conservation efforts may sustain the presence of the breed(s).

Reproductive and production performance of indigenous cattle in Rwanda
From our study, the age at sexual maturity ranged between 18 and 36 months for male Inyambo cattle while for females, the range was narrower though means for the two sexes were close (Table 6).The mean values were much higher than 23.6 and 22.7 months recently reported for Ugandan Ankole cattle (Kugonza et al., 2011a).Average age at first calving (AFC) was 33.8 months and the range of 27-45 months did not vary from 24-45 (Kugonza et al., 2011a) but varied widely from 42-60 (Payne and Wilson, 1999) and Rege (1999) for Ankole cattle in Uganda.These variations between the contemporary studies could be attributed to variations in improvements in general cattle management practices, particularly disease control and feeding, and shifts in breeding objectives from aesthetic (coat colour patterns and horn shapes) to productive traits (fast growth, body weight, and milk yield).Calving interval, a sparring partner of AFC was similar to 12-18 months recently reported elsewhere (Kugonza et al., 2011a) and better than past studies (Payne and Wilson, 1999).Calving percentages (70%) though much lower than values reported for dairy breeds in Rwanda, it was higher for Inyambo compared to other indigenous ecotypes (Table 6).Calvings across a calendar year appear to mainly occur in August-October period (Figure 1).This matches squarely with the main rainfall and cropping season.The implication of this is that cows are able to drop their calves when pasture is available and this supports lactation performance.
In general, our study showed that the mean calving rate of indigenous cattle in Rwanda according to the cattle keepers is 55.8%.A conception rate of 79% in Ankole, Boran and unspecified zebu-type cattle in western Uganda was reported (Trail et al., 1971), and was attributed to abundant feed and water alongside diseases controlled.Under traditional farm conditions but with Tswana cattle in Botswana, Rennie et al. (1976) reported average calving rate of 46.4% incomparable with 74.0% for similar animals on a ranch.Under ranch conditions, higher calving rates would result from animals there being better fed and managed than those from the traditional management system.Elsewhere, White Fulani cattle raised on government ranches in Nigeria had percentage of 67%, incomparable with 34-55% for similar animals raised by local herders (Nuru and Dennis, 1976).We observe therefore that Inkoromoijo and Bashi cattle have low calving rate compared to Inyambo and Inkuku.Low calving rate is due to many a small number of "breedable cows" present.It may also be due to agerelated factors.
On the other hand, regarding pre-weaning calf mortality, Inyambo cattle did not differ from the other Rwandan indigenous ecotypes (Table 6).There was a clear gulf between the ecotypes regarding weaning weights as well as whole lactation yield.Weaning weight was highest in Inyambo, followed by Inkuku, Inkungu, Inkoromoijo and Bashi.The average lactation length was 204 days slightly lower than 216 months (Kugonza et al., 2011a) and 246 months (Twinamatsiko, 2001).Among the Nguni cattle breed of South Africa that is also Sanga but that has largely been intensively undergoing selection (Bester et al., 2003), much higher lactation lengths (245-  (Petersen et al., 2003), and 2.2 kg (Kugonza et al., 2012), for the same breed and production conditions.The yield appears to rise with time and could be a result of deliberate or inadvertent selection for milk yield.When projected over an entire lactation, the resultant values supercede those previously estimated at 550 kg in pastoral areas for Ankole cattle (Kugonza et al., 2012), and appear to match the production (300-1100 kg) of a strain of the Small East African zebu in Bukedi, eastern Uganda reported by Payne and Hodges (1997).Inyambo cows produce milk that is higher in protein, butter fat and minerals than crosses with Friesian (Table 7).However, in terms of yields, crossbreds were superior in all parameters.

Purpose for rearing and record keeping in indigenous cattle in Rwanda
The purpose for rearing cattle in Rwanda is presented in Table 8.Male indigenous cattle are primarily used for income generation (41.7%) when sold and for meat production for the household (37.5%).Female cattle are reared majorly for milk production (45.8%).In all, a total of seven functions are derived from indigenous cattle.From these results, we observe that indigenous cattle continue to be multi-purpose and their valuation has to take this into consideration.Exotic cattle on the other hand predominantly reared for milk (76% of the households).A substantial proportion of households (20%) also derive income from their exotic herd while a very small proportion (4%) use some of their exotic stock for meat, probably the male calves, as has been documented elsewhere (Kugonza et al., 2011a;Nabasirye et al., 2012).Inyambo cattle farmers reported that poor health is the main reason for culling indigenous bulls and cows, as indicated by 37.9% of respondents (for bulls); 20.8% of respondents (for cows); 29.2% of respondents (for young males) and 16% of respondents (for young females).Overall, slow growth was the dominant reason for culling (Table 9) but specifically, bulls are culled mostly because of poor health and old age, cows are culled on the basis of old age and infertility, while both young males and females are culled on the basis of slow growth.
Most of the bulls are sourced from within the cell "umudugudu" (Figure 2), and three quarters of the respondents' source bulls within the district, only 25% sourced outside the district and of these only 5% source bulls outside the country.These statistics are indicative of a likelihood of inbreeding, since majority of farmers reported that the source of the bull is within the cell, Umudugudu.Literacy levels of the cattle keepers were quite high (Table 10), and this enabled the cattle keepers to keep some records on births and health status.

Morphological and metric characteristics of indigenous cattle in Rwanda
The average linear body dimensions of male cattle were higher than for females for different age groups and breed ecotypes (Tables 11 and 12).Parameters for the   top/backline and an evenly sloping rump; bulls additionally have cervico-thoracic (neck) humps.The cattle are spotted in colour but single coat coloured animals are dominant.Horns are white, long and symmetrical, with a large base and proportional to horn length (Figures 3 and 4).The Inkuku is medium in size, and has previously been described as weighing 250-350 kg (cows) and 300-400 kg (bulls) (Felius, 1995), though in or current study, the mean weight was 399 kg (Table 12).Due to their low birth weight, the bulls of this ecotype are useful for breeding to first-calf-heifers of other small breeds.The horns of the adults serve as formidable weapons against any intruders.The Inkuku was previously classified as a short animal that was dominant in high altitude areas (Felius, 1995), with short horns and strong legs to enable  it to adapt to the environment of high altitude (Figure 5).However, in our study, we found animals of this ecotype in the southern midlands, and the lowlands, the spread having possibly been a result of the displacement and migration in the post-1994 nationwide genocide against the Tutsi in Rwanda.
The Inkoromoijo cattle have a very small body, shorter horns and finer boned (Figure 6) than the Ankole type or Sanga breeds (Felius, 1995).The Inkoromoijo have a smaller body frame size, with cows weighing 200-300 kg and bulls weighing 200-330 kg.In this study, we notably found a mean weight of cows of 399 kg.These cattle are predominant located in Eastern province of Rwanda.The Inkungu have much smaller body, than the other types and are characteristically hornless / polled.They have a small size, with cows weighing 230-300 kg and bulls weighing 230-330 kg (Table 12).Anecdotal reports indicate that calves weigh 25-30 kg at birth.The Bashi cattle (Figures 7and 8) have a smaller body, shorter horns, is finer boned than the other Ankole ecotypes (Felius, 1995).The agricultural-based Bashi tribe brought these cattle with them in the 17th century and settled in their present location (Rege and Tawah, 1999) in western province of Rwanda and in Democratic Republic of Congo.In Rwanda they are located in Rubavu and Rusizi districts.Bashi cows weigh 220-300 kg while bulls weigh 200-330 kg.Newborn calves have been reported to weigh 23-25 kg.Whereas Inkungu (Figure 9) were categorised as a separate breed by the farmers in this study, elsewhere, polled cattle are just a component of a breed, for instance among Ankole cattle in Uganda (Kugonza et al., 2011a), this would probably explain why no studies at breed level have focused on this trait exclusively.Figure 10 shows the various coat colour patterns found among cattle in this study.
Table 13 gives the phenotypic correlation estimates between the various traits for female Inyambo cattle.The correlation coefficients between BW and HW, HuL, BL, HL, DW, and LH were positive and different from zero (P<0.05).A high positive correlation (0.98) was also observed between BW and HG; while correlations between BW and HD, and TeL and UD were negative.Generally, BL was negatively correlated to other traits HD, RW, TeL, and UD.Body weight, hearth girth, height at withers, hump length, and dewlap width were all more correlated amongst the group than the rest of the studied traits.
Whereas the differences between the morphometric measurements of male and female cattle are attributable to sexual dimorphism (Mwacharo et al., 2006), that results from hormonal differences between the two sexes at respective ages (Kugonza et al., 2011a), variations between the different ecotypes of Ankole cattle may not be easily attributed to a particular phenomenon.
The measurements for Inyambo cattle in this study are comparable for values published for the Ankole cattle generally (Ndumu et al., 2008) and other Sanga breeds (Payne and Hodges, 1997), but lower than values for  (Mashoko et al., 2007).We did find that Inyambo were the superior ecotype followed by Inkuku, Inkungu and Inkoromoijo in descending order.These variations in morphological descriptions are useful for making distinction between animal strains (Gatesy and Arctander, 2000), evaluation of breeding goals (Zechner et al., 2001), and comparing feeding and production systems as postulated above.The height at withers helps in visually appraising beef cattle classes for show business (Alderson, 1999).Whereas width, girth and body weight measurements are related to muscle and fat deposition which are affected by nutritional status of the cattle, skeletal measurements including height at withers and body length are better indicators of inherent size (Kamalzadeh et al. 1998).The high positive correlation (0.98) observed between BW and HG is similar to values of Kugonza et al. (2011a) for the two traits and affirms their reliability for use as proxy measures for accurate estimation of body weight.Qualitative morphological data on Inyambo cattle showed that majority of the cattle have uniform coat colour pattern (Table 14).Interestingly, the preferred coat colours are different forms brown, namely dark brown, brown, light brown and spotted light brown.The colour interest are largely cultural though previous research found it useful in parentage assignment (Kugonza et al., 2012a) and others postulated an association of body coat colour and pricing of Ankole cattle (Kugonza et al., 2011a;2012b), of which Inyambo are a component.Horn shaping and spacing are also critical traits in enabling cattle to graze in thickets and difficult terrain (Kugonza et al., 2012b).

Conclusions
Ankole cattle show a high productive potential, though their current performance is relatively low, largely because of the sub-optimal management conditions, and production area constraints.Wide variations in management activities and production levels were also observed in the three livestock production systems.However, this variation in the production levels is indicative of potential for improvement under selection and improved husbandry.A set of six traits are good predictors of body weight.It is good to improve and strengthen the current conservation activities; there could be risk of extinction, leading to loss of this genetic material.It is important to note that for an effective cross breeding, pure parent lines have to be maintained, with a selection program for each of them to improve the genetics and maintain the specific traits of that line.So, in case of Rwanda where crossbreeding is encouraged to  increase production through improved genetics, the local breeds lines must be maintained and breeding plans should be designed.In group discussions, all farmers expressed their willingness to participate in and  contribute materially or financially to any program towards a sustainable use and preservation of the indigenous breed which they perceived as hardy and embedded in their culture.We therefore argue that strategies for its sustainable use and conservation should consist of simultaneously improving general herd management practices, organizing farmers and involving them in participatory breed improvement programs.

RECOMMENDATIONS
There is need to strengthen the conservation efforts for the various ecotypes of indigenous cattle of Rwanda.
Particularly, the Bashi cattle appear to be severely substituted by the Holstein Friesian and is almost absent breed in general.Extension services for cattle farming should be strengthened through increasing the number of public and private service providers through affirmative action.

Figure 3 .
Figure 3. Inyambo cows characterized by large body frame with long horns.

Figure 4 .
Figure 4.An Inyambo bull characterized by dark brown coat colour, a large body frame and long horns.

Figure 5 .
Figure 5. Inkuku cows have a small body framework with medium sized horns.

Figure 6 .
Figure 6.An Inkoromoijo cow in the foreground.

Figure 7 .
Figure 7.The Bashi cattle in a woodland area.

Figure 8 .
Figure 8.The Bashi cattle on the open range.

Figure 10 .
Figure 10.Body coat colours and patterns in indigenous cattle of Rwanda.

Table 1 .
Study sub-locations, altitude zones and management systems.

Table 2 .
Number of cattle by age category, sex and breed used in the characterisation study.

Table 3 .
Household membership, cattle herd and small stock flock structure.

Table 4 .
Major farming activities, household income sources and species of livestock reared.
*The responses on major income sources were non-exclusive.

Table 5 .
Grazing management and herd dynamics for Inyambo cattle in previous year.
*Transferred from another herd; ϕ HH means household.

Table 6 .
Productivity and reproductive parameters of indigenous cattle of Rwanda.
*Projected from daily milk yield and lactation length; For the Inkuku, Inkungu, Inkoromoijo and Bashi cattle, only mean values are given.
Leeuw and Wilson, 1987)orded(Rege et al. 2001).The mean milk off take for Inyambo cows in the current study is quite substantial though with a wide range.At 3.6 kg, the value varies widely from 1.08 kg (DeLeeuw and Wilson, 1987), 1.48 kg

Table 8 .
Purpose of rearing Inyambo cattle among cattle farmers.

Table 9 .
Reason for culling the Inyambo cattle.
Figure 2. Source of breeding bulls.

Table 10 .
Literacy level and cattle record keeping among Inyambo cattle farmers.
Inyambo were superior to those of Inkuku and Inkoromoijo at different ages.For bulls, it was difficult to measure them beyond weaning age.Mature body weight of Inyambo cows was 482 kg whereas the bulls had an average weight of 647.0 kg.In appearance, these cattle are elegant, well-bred, and graceful, have a straight

Table 12 .
Body morphometrics of Inkungu and Bashi cattle in Rwanda.

Table 13 .
Correlation matrix for body measurements of adult Inyambo cows.