Species abundance, food preference and nutritive value of goat diets in the semi-arid lands of east-central Kenya

N.K. Osolo, J.N. Kinuthia, C.K. Gachuiri, A.M. Okeyo,
M.M. Wanyoike and M. Okomo

University of Nairobi, Faculty of Veterinary Medicine, P.O. Box 29053, Nairobi, Kenya

Abstract
Introduction
Materials and methods
Results and discussion
References

Abstract

The key forage species and nutrient composition of forage species preferred by free-ranging East African goats in rangelands of eastern-central Kenya were determined. One hundred and five forage species were identified including 30 tree, 29 shrub, 20 grass and 25 forte species. Shrubs and short trees were the most preferred forage species and thus formed the bulk of the goat diets. Ironically the forage species most preferred by the goats were not necessarily the most abundant on the ground, indicating that ease of accessibility and palatability may be important determinants for choice of a forage specie. That grasses and fortes were the least preferred could be largely explained by the fact that goats are naturally browsers.

Introduction

Raising of livestock by pastoralists, agro-pastoralists or commercial ranchers is the single most important economic activity in the arid and semi-arid lands (ASAL) of Kenya. Over two-thirds of Kenya (approximately 492,100 km²) is classified as ASAL. More than half of this area receives less than 370 mm of variable rainfall annually and is further characterised by variability in pasture and water availability. During the peak of the rainy season, feed resources available to animals are high in quality and quantity and animals gain weight fairly fast. Conversely, at the peak of dry spell, available forage resources are too low in quality and quantity and animals inevitably lose weight. Important domestic animals in ASAL include, cattle, sheep, goats and camels.

Recent increases in the human population due to better living standards, have been accompanied by increases in total livestock numbers in the ASAL, especially among pastoralists who, to a large extent, depend on these animals for their subsistence. Increase in human population in high potential areas has also resulted in migration of people to the ASAL in search of arable land. In addition, among the indigenous communities, there is a growing tendency towards a more cash-oriented economy as evidenced by subdivision of communal lands into small individual parcels and acquisition of title deeds. Thus, families are increasingly becoming sedentary, but on land units too small to meet their food needs. There is therefore overstocking resulting in range degradation especially around settlements and cultivated areas.

Even with individualisation of land ownership, diminishing household land units and the terminal breakdown of traditional communal grazing systems, livestock production remains the mainstream economic activity for these areas. In order to make meaningful recommendations to land owners on proper management of these lands, especially under the currently emerging land use/ownership paradigms, range managers ought to have a thorough knowledge of the most suitable kinds of animals, their nutrient requirements, forage preferences and quality of available forage.

Information on food habits of pastoral livestock in the East African ASAL is scarce. Livestock select their diet from a complex mixture of plant species and/or plant parts (Arnold 1960; Arnold 1962). Why an animal selects one plant and not another is still a puzzle. Choice varies widely and seems to depend on availability. Diet selection may modify the species composition of the pastures in the long run (Illius 1986) with the potential danger of lowering the quality of the diet. Continued grazing pressure increases the number of unpreferred plant species at the expense of the preferred species (Tadingar, personal communication, 1984). Due to wide temporal and spatial variation in forage preference and nutrient composition of diets selected by different kinds of ruminants, research information generated from a given area and kind or class of animals, has limited inference (space) and should not be applied to too wide an area. It is therefore, important to establish local food habits and nutrient composition data bases for the various types of domestic and wild animals utilising particular range types.

The main purpose of this study was to determine the key forage species and nutrient composition of diets selected by free-ranging East African goats in rangelands of eastern-central, Kenya.

Materials and methods

Study area and procedures

The study was conducted at the University of Nairobi's Machang'a Field Station situated approximately 150 km east of Nairobi and next to Kamburu Dam on the eastern bank of upper Tana river. The station is approximately 500 ha in area and lies at 1°S latitude, 37°E longitude and about 700 m altitude. However, the study was limited to a 20 ha area fronting the Embu-Kitui main road. The area gently rises from the Kamburu Dam to the top of the ridge.

The mean annual rainfall is 650 mm coming in two seasons - March through May, short rainy season and October through December, long rains. Soils on the station are developed from undifferentiated types of gneisses of Basement System. They are well drained, shallow, dark red to yellowish red, stony loamy sand to clay (Kenya Soil Survey, Misc. Soil Paper M 19, 2nd impr. Nairobi 1979). Machang'a vegetation is described as bushed shortgrass savannah with buffer grass (Cenchrus ciliaris) and horsetail grass (Chloris roxburghiana) predominant. Saltbush is a common palatable shrub suitable for re-establishing pasture on overgrazed and eroded places (East African Livestock Survey Atlas, 1967). Recommended stocking is about three hectares per livestock unit (LU).

Species composition and relative abundance

At the beginning of the November/December wet (green) season of 1993, botanical composition and relative abundance of both herbaceous and woody plant species layers were determined. This involved the use of line intercept method, where 10 transects, 100 m long by 200 m apart, were laid across the slope and all herbaceous plant species intercepted by the tape identified. Herbaceous layer determination involved the quadrat method, where ten 1 m² quadrats were laid at equal intervals along each transect. Each quadrat was further subdivided into four quarters and different plant species occurring in each quarter identified, enumerated and recorded. For trees and shrubs above 2 m high, 10 m x 10 m quadrats were used and different species occuring within each quadrat identified, counted and recorded. The relative abundance for each plant species was calculated as follows:

Relative abundance of sp A= Number of times sp A occurred/No of times all sp occurred

To convert number of plants per m² into number of plants per ha, the density value is multiplied by 10,000 m².

Number of bites and time spent on individual species

The method used to study forage preference involved close observation of randomly selected animals (one at a time) as they grazed with the others. A base flock of 45 East African goats was used from which one animal was selected for the bite count study. The selected goat was closely followed and observed by two enumerators who recorded the number of bites and time spent on each plant species as the goat grazed or browsed. A bite was taken as the act of breaking off or picking up a piece of forage. The time spent by the animal feeding on one individual plant without walking one full step was estimated in seconds and recorded. The animal was observed for 30 minutes beginning at 0930 hours.

The flock was grazed in different areas and a different goat was randomly selected for observation so that different goats were used on the five grazing units. The five goats selected during the first round of grazing rotation qualified as experimental animals for subsequent replicates. This observation procedure was repeated for as long as there was no noticeable change in the vegetation, i.e. until the rains began if it was a dry-season grazing, or until the dry season set in if it was green-season grazing.

Forage digestibility and chemical composition

Plant material representative of that selected by the goats was hand-clipped, sun dried for three days and then oven dried at 40°C for four days. The dry forage samples were ground in a Wiley Mill using a I mm screen. The material was later chemically analysed for NDF, ADF (Van Soest and Wine 1967); N P. Ca, Mg, K, Na, Zn, Cu. and Fe (AOAC 1975). Digestibility estimates using the nylon-bag technique (Ørskov et al 1980) for 48 hours incubation was also conducted.

Results and discussion

Vegetation: Botanical composition

Tables la and b give an inventory of the plant species occurring in the study area, and their relative density (plants/ha). One hundred and five plant species were identified comprising, 30 trees, 29 shrubs, 20 grasses and 25 fortes. Since the survey was conducted soon after the onset of the short rains, a large number of tree and shrub seedlings was observed.

The most dominant tree species within the study area was Acacia brevispica. Others included Combretum zeheri, Cammiphora riparia, Terminalia brownii, Grewia villosa, Acacia nilotica, Boswelia sp, Lannea flocesca and Commiphora boviniana in that order. The most dominant shrub was Hermania alhiensis. Others, in order of importance, were Grewia bicolor, Gnidia latifolia, Premna hilderbrandtii, Dichrostachys cinerea, Acalypha fruticosa and Maytenus putherlickioides. Of the grasses, annuals dominated the site due to their opportunistic nature. The most dominant grass species was Aristida keniensis, followed by Eragrostis caespitosa and Latipes senegalense. Most of the perennial grasses had not come up. Among the fortes, Ocimum sp was most dominant, followed by Rhoicissus tridentata, Pentas sp, Stylosanthes aspera and Acanthospernum hispidum.

Table 1a. Average plant population density (plants/ha) of trees and shrubs in the study area.

Trees	Plants/ha	Shrubs	Plants/ha
Acacia etbaica	10	Withania somnifera	10
Acacia brevispica	65	Triumfetta s p	40
Lannea triphylla	35	Grewia bicolor	80
Combretum moue	35	Sida ovata	45
Walbergia sp	10	Combretum aculeatum	45
Grewia villosa	40	Premna hilderbrandtii	75
Acacia nilotica	40	Indogofera schimperii	15
Acacia mellifera	20	Indigofera lupatana	5
Commiphora riparia	50	Acalypha fruticosa	70
Boswelia sp	40	Grewia similis	35
Commiphora africana	35	Hibiscus calyphylla	15
Zanthoxylum chalybea	5	Aspangus racenosus	40
Terminalia brownii	50	Hermania alhiensis	220
Acacia tortilis	5	Ormocarpum trichocarpa	50
Combretum zeheri	55	Cassia singueana	15
Acacia hockii	25	Dichrostachys cinerea	75
Steganotaenia araliaceae	5	Thumbergia holstii	10
Ochna inermis	10	Tephrosia villosa	20
Albizia amara	10	Ipomoea kituensis	5
Lannea floccosa	40	Maytenus putherlickioides	70
Commiphora boviniana	40	Ormocarpum kirkii	5
Commiphora campestris	10	Gnidia latifolia	80
Terminalia spinosa	5	Hoslundia opposita	25
Dalbergia melanoxylon	10	Indigofera clifordiana	15
Euphobia nyikae	25	Crotolaria gooformis	40
Commiphora schimperii	5	Croton dichogamus	5
Makhamia sp	15	Rutya fruticosa	5
Entada abyssinia	5	Hibiscus micranthus	20
Lonchocarpus sp	10	Zimania americana	5
Cordia gharaf	5

Table 1b. Average plant population density (plants/ha) of grasses and forbs in the study area.

Grasses	Plants/ha	Forbs	Plants/ha
Setaria verticilata	t*	Rhoicissus trindentata	5
Dactyloctenium aegyptica	2	Ocimum sp	80
Brachiaria reptans	1	Pentas sp	5
Aristida keniensis	6	Stylosanthes fruiticosa	5
Enteropogon macrostachys	2	Achranthes aspera	3
Eragrostis superba	1	Solanum incanum	3
Chloris virgata	t	Alternthera pungens	2
Eragrostis caespitosa	3	Acanthospernum hispidum	4
Latipes senegalense	3	Commelina benglhelensis	t
Heteropogon controtus	t	Tribulus terestris	1
Michrochloa kunthii	2	Pupalia lappaceae	1
Themeda triandra	t	Convolvulus favinosus	1
Digitaria macroblephera	2	Oxygonum sinuatum	t
Chlorisroxburghiana	t	Indigofera spicata	t
Digitaria gazensis	1	Ocimum bacilicum	1
Latipes senegalenses	t	Tephrosia pumila	t
Aristida adoensis	1	Blepharis hildebrandtii	t
Digitaria falutina	1	Phyllanthus maderaspatens	t
Tragus berteronianus	1	Barleria proxima	t
Rhynchelitrum roseum	t	Asystasia schimperi	t
Aristida sp	t	Hibiscus micranthus	t
		Rhoicissus triandenta	t
		Zornia glachidiata	1
		Bidens pillosa	1
		Ipomoea oenothrae	t

* t = trace amounts (<0.05 plants/ha).

Forage preference

Shrubs and short trees were the most preferred forage plants and hence formed the bulk of the goat diets (Table 2). The five most preferred shrubs were Maytenus putterlickioides, Premna hilderbrandtii, Hoslundia opposita, Dichrostachys cinerea and Combretum aculeatum in that order. However, these plants were not the most abundant on the ground suggesting that the plants/grazable plant parts were possibly both more accessible and palatable and therefore more seriously sought after by the goats. Only Premna sp showed a close relationship between its relative preference and its abundance on the ground - third and second, respectively.

Table 2. Mean number of bites and time (in seconds) spent on various forage species within the study area.

Forage species	Mean number of bites	Mean time spent
Maytenus putterlickioides (s*)	31	44
Ochna inermist (t)	24	36
Premna hildebrandtii (s)	23	30
Hoslundia opposita (s)	21	31
Dichrostachyus cinereas (s)	18	24
Combretum aculeatums (s)	18	20
Acacia hockii (t)	15	19
Grewia bicolor (s)	14	22
Acacia mellifera (t)	14	17
Thumbergia holstii (s)	13	16
Acalypha fruticosa (s)	13	16
Barleria proxima (g) f	13	18
Combretum moue (t)	11	20
Ocimum spp (f)	10	13
Grass (bulked)	8	10
Boswelia hilderbrandtii (t)	7	10
Hermania alhiensis (s)	4	11

*s = shrubs,
t = tree,
g = grass,
f = forte.

Table 3. The correlation and regression coefficients between plant species preference and nutrient components.

Nutrient component	Correlation	Simple regression	Partial (multiple regression
	coefficient	coefficient	coefficient)
ADF (%)*	0.3400	0.35 (0.27)	-0.23 (0.59)
NDF (%)	0.0001	0.16 (0.21)	0.25 (0.49)
IVDMD (%)	-0.2700	-0.12 (0.11)	-0.13 (0.21)
Nitrogen (N)	0.3000	0.30 (0.29)	0.55 (0.41)
Phosphorus (P)	-0.0500	-1.83 (1.83)	-3.16 (4.28)
Sodium (Na)	0.3400	0.34 (0.03)	0.06 (0.03)

* ADF = acid-detergent fibre;
NDF = neutral-detergent fibre;
IVDMD = in vitro dry-matter digestibility.

Among the trees, Ochna inermis and Acacia sp were the two most preferred plants. Ochna inermis ranked second to the most preferred shrub, Maytenus putterlickioides. Ochna sp, however, ranked low in availability on the ground. Grasses and fortes were least preferred which is because goats are naturally browsers, and these plants were least abundant on the study area.

Table 3 represents correlation and regression of the number of bites (preference) on nutrient components (ADF, NDF, IVDMD, N. P. and Na). There was a positive and significant correlation between ADF, N and Na with number of bites; P showed a significant negative correlation.

ADF, N and Na had a significant positive causal (regression coefficient) effect on preference. NDF, IVDMD and P had no significant causal relationship with preference although P and digestibility had insignificant negative regression coefficients (-1.83 and -0.12, respectively) on preference. It can therefore be concluded that the different plant species were more preferred due to relative contents of ADF, N or Na and less preferred due to relative P contents (Tables 3 and 4).

A multiple regression analysis showed that the most important components that influenced preference in this study were N (0.55), and P (-3.16). NDF (0.25) and ADF (0.23) were less important due to large standard errors associated with the estimates. Conversely, although the partial regression coefficients of digestibility (-0.13) and phosphorus (3.16) were negative the estimates were insignificant due to large standard errors. Work is continuing in this area and with more data, a clearer picture of these relationships will become evident since it is possible that other components of the diets such as saponins and tannins that were not considered in this study could have played a major role in determining the relative dietary preference.

Table 4. Nutritive value of some of the forage plants preferred by goats at Machang'a.

Plant species	ADF	ADF	Ca	Mg	K	Na	Zn	P	N
	%	%	(g/kg)	%	(mg/kg)	(mg/kg)	(mg/kg)
Premna hildebrandtii	26.73	35.80	8.15	1.6	15.7	272	67	2.3	23.4
Grass (bulked)	27.96	35.58	4.5	1.8	17.1	116	34	2.5	14.6
Indigofera spicata	28.98	35.53	55.3	6.2	13.3	61	32	-	-
Acacia ataxanthe	23.63	36.90	20.0	3.8	7.6	75	54	-	-
Boswelia hilderbrandtii	30.43	42.97	1.7	6.8	7.6	95	54	1.0	10.8
Aspiilia mossambiscensis	30.43	42.97	32.1	4.3	24.9	48	64	1.6	29.8
Maytenus putterlickioides	38.97	46.89	15.5	4.6	7.5	163	31	0.8	18.2
Combretum sp (bulk)	30.93	36.92	8.2	3.1	10.4	48	26	3.0	21.0
Crotolaria goodformis	20.39	36.14	10.3	2.5	17.6	34	64	-	-
Acacia hockii	22.21	36.21	11.2	1.9	8.7	-	28	0.8	13.3
Tephrosia villosa	35.48	53.28	11.4	2.5	10.4	55	41	-	-
Uchna inermis	34.96	55.91	19.8	3.4	4.9	20	23	0.8	25.1
Dichrostachys cineria	44.24	55.10	12.2	2.8	7.2	109	25	-	-
Grewia bicolor	36.50	59.92	19.8	3.9	11.5	41	20	2.0	21.9
Stephania abyssinica	31.25	49.20	11.7	2.5	20.5	41	50	-	-
Lantana camara	24.62	56.49	19.7	6.7	18.8	48	38	-	-
Dalbergia melanoxylon	38.15	67.47	14.7	2.9	14.5	82	12	-	-
Acacia mellifera	28.06	51.71	17.6	3.6	12.6	129	28	2.8	27.6
Acacia tortilis	46.44	52.39	12.8	2.9	12.6	88	24	-	-
Combretum aculeatum	27.64	48.21	15.1	2.3	11.5	82	73	3.0	21.0
Ocimum sp	28.15	44.53	16.1	4.6	21.6	55	61	1.8	25.9
Barleria proxima	39.63	53.93	39.5	5.9	10.1	109	35	1.8	20.5

ADF = acid-detergent fibre.

References

AOAC (Association of Official Analytical Chemists). 1975. Official Methods of Analysis. 12th edition. AOAC, Washington, DC, USA.

Arnold G.W. 1960. Selective grazing by sheep of two forage species at different stages of growth. Australian Journal of Agricultural Research 11:1027-1033.

Arnold G.W. 1962. Factors within plant association affecting the behavior and performance of grazing animals. In: Crisp D.J. (ed), Grazing in Marine and Terrestrial Environments. pp. 133-184.

Forage behavior and diet selection. In: Gudmundsson O. (ed), Grazing Research at Northern Latitudes. Plenum Press, New York, USA.

Ørskov E.R., De Hovell F.D. and Mould F. 1980. The use of the nylon bag technique for the evaluation of feedstuffs Tropical Animal Production 5:195-213.