Productivity of oversown natural pastures in northern Tanzania

M.L. Kusekwa¹, S.N. Bitende² and M.D. Ngowi¹

¹ Tanzania Livestock Research Organisation, Livestock Production Research Institute, Mpwapwa, Tanzania.

² Tanzania Livestock Research Organisation, Livestock Research Centre, Tanga, Tanzania.

Abstract
Introduction
Materials and methods
Results
Discussion
Conclusion
Acknowledgement
References

Abstract

Natural pastures are the cheapest and main source of feed for the ruminant livestock population in Tanzania. These pastures occupy about 60 million hectares of Tanzania mainland's approximately 87 million hectares. The majority of the cattle, goats and sheep are kept by small holder livestock keepers in rural areas where communal grazing is commonly practiced (Anon, 1983). Herbage production of these pastures is contained by overgrazing, poor species composition and general land degradation.

Oversowing of natural pastures with improved forage species, particularly legumes, improves the production and quality of resulting pastures. At Tengeru, Arusha it was found that natural pastures oversown with Desmodium intortum produced an average of 15.6 tonnes DM/ha annually as compared to annual yield of 13.0 of tonnes DM/ha from natural pastures alone. Liveweight gains of heifers grazing oversown natural pastures were, however, only slightly higher (10%) than those of heifers on natural pasture alone.

This paper discusses the potentials of oversown natural pastures in the improvement of livestock productivity with special reference to work done at Tengeru, Arusha, northern Tanzania.

Introduction

Natural pastures are important in Tanzania in that they support most of the ruminant livestock population. However, their productivity is low due to the inherent low production and quality of the component species (French, 1957; Calo 1976; Mwakatundu, 1977).

Natural pastures can be improved through employing the following ways:

i) Improvement of the management and utilisation of the existing natural pastures.

ii) Replacement of natural pasture species with improved ones and

iii) The use of the combination approach involving oversowing of natural pastures with improved species, particularly legumes.

Improvement of management and utilisation of natural pastures is in itself constrained by the inherent low productivity and quality of herbage of the existing plant species (Calo, 1976). Replacing them with improved species by cultivation and seeding is unlikely to be adopted due to the high costs involved. A more realistic approach would appear to be the employment of oversowing techniques for natural pastures development. Limited experience in Tanzania has shown that desired species could be successfully introduced into the natural pastures with minimum operations such as hard grazing, burning, use of herbicides and minimum cultivations (Northwood and Macartney, 1969; Lane and Lwoga, 1978; Kusekwa, 1982; Lwoga, 1983). The work by Anderson and Naveh (1968) and Naveh (1967) in northern Tanzania, that reported by Lane and Lwoga (1978) and that of Rukanda and Lwoga (1981) in Morogoro have suggested that a number of legumes could be grown in these areas. Therefore it was a matter of importance to carry out studies on how to successfully establish these species in natural pastures by techniques that employed oversowing methods.

This study has endeavoured to throw some light on oversowing techniques to be used, what species are most suited for oversowing performance of oversown natural pastures under grazing and the growth of animals grazing these pastures.

The objectives of this study were:

(a) To determine the best method of seedbed preparation for the establishment of pasture legumes by oversowing.

(b) To determine the role of phosphatic fertilizers in the productivity of oversown natural pastures, and

(c) To determine the most suitable legume species for oversowing purposes and their role in increasing the productivity of natural pastures in terms of herbage and animal production.

Materials and methods

Grazing Experiment
Stocking Rates
Experimental design
Measurements
Botanical composition
Dry matter production
Liveweight gains measurements

Plot and grazing experiments were conducted from April, 1979 to December, 1983 at the Livestock Training Institute, Tengeru, in Arusha, Northern Tanzania. Detailed description of the research site and the methodology used for the plot experiment are given in earlier reports (Kusekwa and Lwoga, 1986; Kusekwa, 1988).

Grazing Experiment

Two legumes, out of the 9 tested for suitability for oversowing in the plot experiment, Macroptilium atropurpureum (Siratro) and Desmodium intortum (Greenleaf desmodium) were selected and oversown into paddocks after the natural pastures had been hard grazed. The oversown pastures were given one year for establishment before grazing was introduced.

At grazing, two stocking rates were used and the basis for their selection was the recommendations by Pratt et al (1966) that the carrying capacity for Ecological Zone II (i.e. Humid to dry subhumid) was 1-2.5 ha/stock unit. Yearling dairy heifers grazed the pastures for six months of the year during the growing season and part of the dry season. The experiment lasted for 3 years (1981, 1982 and 1983) and, each year, a new batch of heifers was used.

Pasture types (treatments):

(a) Siratro oversown natural pastures (Siratro)
(b) Greenleaf desmodium oversown natural pastures (Greenleaf desmodium)
(c) Natural pasture alone (control).

Stocking Rates

Only two stocking rates were used instead of three or more as generally recommended (Mannetje et al 1976) because land and animals available for experiment were limited.

(a) High stocking rate (SR₁) = 2 animal units/ha or 4 heifers/ha (liveweight ranged from 120-150 kg).
(b) Low stocking rate (SR₂) = 1 animal unit/ha or 2 heifers/ha.

Experimental design

The randomized block design was used in which 12 paddocks made up of all combinations between the three pasture treatments and 2 stocking rates were replicated twice. Three heifers were allocated to each paddock. Thus the paddocks for the high were 0.75 ha each and those for the low stocking rate were 1.5 ha each.

Measurements

Various measurements were taken during the course of the study which included soil tests, botanical composition, dry matter production, pasture quality and liveweight gains.

Botanical composition

The Dry Weight Rank (DWR) method as described by Tothill et al (1978). The rank BOTANAL computational package as described by Tothill et al (1978).

Dry matter production

Dry matter production was determined by placing cages in the paddocks to exclude grazing animals. The growth of the pastures was then monitored, that within and outside the cages, by cutting regularly samples for dry matter production assessement.

Liveweight gains measurements

The animals were weighed twice a week during the entire grazing period. Weighing was done in the mornings, from about 7.00 hrs to about 9.00 hrs. The weighing of the animals was arranged such that it coincided with spraying for tick control (on Tuesdays and Fridays).

Results

Annual dry-matter yield
Presentation yield (kg/ha) during the grazing period.
Botanical composition of pastures
Pasture quality
Liveweight gains (LWG) of grazing animals

In this paper, results showing the effects of oversown pasture legumes on natural pasture and animal productivity, dry-matter yield, botanical composition and liveweight gains, will be given. Results on pasture legumes suitability for oversowing, establishment, persistence (survival), dry matter production and responses to phosphatic fertilizer application are reported elsewhere (Kusekwa and Lwoga, 1986, Kusekwa, 1988).

Annual dry-matter yield

The annual DM yield (kg/ha) was not significantly affected by oversowing with legumes as shown in Table 1. However, the 1983 DM yield (11,626 kg/ha) was significantly lower than those of 1981 and 1982 (16,007 and 15,182 kg/ha, respectively). This represented decreases in DM yield of 27% between 1981 and 1983 and 23% between 1982 and 1983. Within pasture treatments the decline in DM yields in 1983 was lower than that for 1981 and 1982 (P<0.05) except for the control pasture where the 1983 DM yield was significantly lower (P<0.05) than that of 1981. However, it was in the control pasture where the decrease in DM yield was largest, 33% as compared to 27% and 21% in the Siratro and Greenleaf desmodium pastures respectively.

Table 1. Effects of oversown pasture legumes on annual dry-matter yield (kg/ha)

LSD (0.05)                     NS         NS      NS      NS       
S.E.                              2596     2901   1193  1350.9

Table 2. Effects of stocking rate on annual dry-matter yield (kg/ha) SR₁ = High Stocking rate, SR₂ = Low Stocking rate.

LSD (0.05)       NS     NS       NS      NS
S.E.                 670.0 841.0  975.0  388.1

Table 3. Overall botanical composition (%) of an oversown natural pasture over three years (1981-1983).

Table 2 shows that stocking rate had no significant effects on DM yield. However, DM yield declined with time: at SR₁ (high stocking rate) there was a decline of 30% between 1981 and 1983. Similarly at SR₂ (low stocking rate) there was a decline of 24% between 1981 and 1983. Both these declines in DM yield were significant (P<0.05).

Presentation yield (kg/ha) during the grazing period.

Presentation yield or available DM during the grazing period, determined by sampling at the beginning, middle and end of each grazing period (H₁, H₂ and H₃ respectively) did not differ significantly between pasture treatments. However, the mean DM yield for sampling dates were significantly different (P<0.05). Mean DM yield at H₃ (4,995 kg/ha) was lower (P<0.05) than that at H₁ (10,856 kg/ha) a reduction of 54% in available DM between these sampling dates.

Stocking rate had no significant effect on presentation yield (available DM) at the various sampling stages during the grazing periods. The declines in available DM between H₁ and H₃ were, however, significant (P<0.05). At SR₁ the reduction in yield was 58% while at SR₂ it was 49%.

Botanical composition of pastures

Table 3 shows that pasture component species, grasses, sown legumes, volunteer legumes sod weeds, did not differ significantly between years. However, in Table 4, all the pasture component except volunteer legumes changed with years. The contact of the sown legume, Siratro in 1983 (2%) was significantly lower than that of 1982 (9.8%) (P<0.05). In the Greenleaf desmodium pastures in Table 5 sown legume and the weed contents differed significantly between years (P<0.05). Greenleaf desmodium content in 1983 (33.6%) was higher than in 1982 (16.7%) while weed content in 1983 (4.0%) was lower than in 1981 and 1982 (7.1% and 7.3% respectively).

Pasture quality

The crude protein contents and digestibility coefficients of leaf herbage from the pastures declined significantly (P<0.05) with time during the grazing period. In the Siratro pasture, CP content dropped from 14% at H₁ to 10.8% at H₃ and digestibility dropped from 66.8% to 60% at H₁ and H₃ respectively. In the Greenleaf desmodium pasture CP content declined from 15.6% at H₁ to 11.2% at H₃ while digestibility declined from 64.3% at H₁ to 55.9% at H₃. In the control pasture, CP content varied between 12.0% at H₁ and 7.3% at H₃ while digestibility was 62.3% at H₁ and 50.3% at H₃. The control treatment, natural pasture alone, tended to drop in quality faster than the Siratro and Greenleaf desmodium pastures. This was indicative of a general improvement: in the quality of herbage of legume oversown natural pastures.

Table 4. Overall botanical composition (%) of a Siratro oversown natural pasture over three years (1981-1983).

Table 5: Overall botanical composition (%) of Greenleaf desmodium oversown natural pasture over three years (1981-1983).

Liveweight gains (LWG) of grazing animals

The LWG per head of heifers was not significantly affected by pasture treatments but was influenced significantly by stocking rates (P<0.05). At SR₁, LWG per head was 56.0 kg and at SR₂ it was 77.2 kg, working out to the equivalent of 311 gm and 429 gm LWG per head per day at SR₁ and SR₂ respectively. Although pasture treatments did not affect significantly LWG per head of the heifers, there was an overall improvement in the growth of the animal grazing oversown natural pastures as shown in Figure 1 and Appendix Figures 1-3.

LWG/ha of heifers during the 6-month grazing period was significantly affected by pasture. At SR₁ and SR₂ LWG/ha was higher, 224.1 kg than at SR₂, 154.3 kg. This represented a LWG/ha/day of 1245 gm and 857 gm at SR₁ and SR₂ respectively. Figure 2 and Appendix Figures 4-6 show the relationship between LWG/head and LWG/ha at the two stocking rates.

Discussion

Dry-matter (DM) yield
Presentation yield
Botanical composition
Pasture quality
Liveweight gains
Relevance of oversowing to pasture improvement in Tanzania

Dry-matter (DM) yield

The annual dry-matter yield of the natural pasture was not significantly affected by legume oversowing (Table 1). However, legume oversown natural pastures in particular, Greenleaf desmodium, consistently produced more DM than the Siratro and control pastures. Clatworthy (1984), in Zimbabwe, observed increased DM production of a natural pasture oversown Desmodium uncinatum (Silverleaf desmodium). Other workers have reported increases in pasture production when suitable pasture legumes were successfully incorporated (Walker, 1969; Stobbs, 1969 and Lwoga, 1983).

The general decline in DM production over the three years could be attributed to effects of defoliation and hence fertility decline and depletion of food reserves for regrowth. It is shown that DM yield declined the least on the Greenleaf desmodium pasture, 21.6% as compared to 27.7% and 33.4% on the Siratro and control pastures respectively. This follows closely the legume content in the pasture in that, the higher the legume content in the pasture the smaller the decline, a function of fertility build-up through N-fixation by the legume component. The abnormally low rainfall in 1983 (Appendix Table 1) is thought to have further depressed DM yield that year.

Presentation yield

Presentation yield at the various sampling dates was not significantly affected by oversowing with legumes although Greenleaf desmodium pastures had consistently more dry-matter yield at all sampling dates during the grazing period. Keya et al (1971) observed a 28% and 31% increased in DM yield in a natural pasture oversown with Desmodium intortum and D. uncinatum respectively. Presentation yield declined at each sampling date and decline was more severe at SR₁ than at SR₂. This could be attributed to faster removal of herbage than was being replaced by regrowth during the grazing period while the differences between SR₁ and SR₂ could be attributed to more severe defoliation at SR₁. Other workers have reported declines in DM yields with increase in stocking rates caused mainly by increase in stocking rates caused mainly by increase in bare ground and unproductive species (Jones et al, 1984; Mannetje, 1984). However, in this study, there was no evidence of pasture deterioration caused by grazing pressure exerted by SR₁ and SR₂.

Figure 1. Monthly cummulative liveweight gains of heifers kg/head.

Mean of 3 years (1981-1983)

Figure 2. Monthly cummulative liveweight gains of heifers kg/head and per hectare.

Mean of 3 years (1981-1983)

Legend:

BR₁a = LWG/head at the high stocking rate
BR₁b = LWG/ha at the high stocking rate
BR₂a = LWG/head at the low stocking rate
BR₂b = LWG/ha at the low stocking rate

Botanical composition

Botanical composition was not influenced significantly (P>0.05) by either pasture treatment or stocking rate. It has been reported that high stocking rates tended to encourage weed invasion and depletion of desirable species (Humphreys, 1981). In this study, both SR₁ and SR₂ were significantly lenient to maintain stable pastures.

Botanical composition did not change significantly (P>0.05) between sampling dates or between years. However, the Siratro pastures dropped significantly (P<0.05) in their legume content (from 2.8% in 1982 to 2.0% in 1983). This suggests that Siratro was being selectively grazed and that regeneration through seed setting was hampered (Jones, 1981; Tothill and Jones, 1977).

Pasture quality

Pasture quality in terms of CP content and digestibility was found to be higher in legume oversown natural pastures than in natural pastures alone. The decline in quality was also slower when initial values were higher (Tothill, 1986). This points to the usefulness of grass pastures as overall feed quality is improved. Oversowing of forage legumes into natural pastures has also been reported to increase the quality of the resultant herbage where Macroptilium atropureum and Stylosanthes guianensis were oversown in Themeda. In vitro DM digestibility and CP content were 55.7% and 9.1% respectively on oversown natural grasslands as compared to 51.9% digestibility and 8.1% CP on natural grassland alone (Rukanda and Lwoga, 1981).

Liveweight gains

The liveweight gains of heifers was not significantly influenced (P>0.05) by pasture treatments. However, the liveweight gains, per head and per hectare, on the Greenleaf desmodium were consistently higher than those on the Siratro and control pastures. The lack of a significant difference in the LWG's between pasture treatments is in agreement with observations by other workers who found that, in the wet season, animals performed similarly on natural pastures and on improved pastures (Mannetje, 1984; Tothill, 1985 pers. comm; Tothill, 1986). Feed quality improvement resulted in increases in LWG's and that the quality factor was more evident in the dry season. In this study quality of pasture in terms of CP did not fall below the suggested critical level of 6.5% CP (Jones et al, 1984).

Stocking rate influenced significantly (P<0.05) LWG per head and per hectare with the higher stocking rate giving higher LUG's/ha than the low stocking rate while the pastures remained fairly stable. Intermittent 6-month continuous grazing of oversown natural pastures appear to have been appropriate at both SR₁ and SR₂ evidenced by the relatively stable botanical composition. The higher DM yield, presentation yield, legume content and leaf CP content on the Greenleaf desmodium pasture than on the Siratro and control pastures reflected the overall better LWG's of heifers on this pasture.

Relevance of oversowing to pasture improvement in Tanzania

It is our view that this technology has a role to play in the development of forage resources in Tanzania. The majority of the reasons advanced for poor adoption of innovations by small-scale farmers also apply with regards to oversowing. However, we specifically consider the following factors to be important: - Oversowing was best suited to systems where some form of land ownership existed, namely commercial holdings under governmental, parastatal or private control. Here the technology has been and is being employed for pasture development. Small-scale farmers could readily adopt this technology (all other things being equal) through setting aside a piece of land, on communal land, for development. The rule practiced here was "use the land, gain it and vice versa" (Anon, 1986).

Necessary inputs such as seeds, appropriate machinery and management were in short supply. This was likely to slow the adoption process.

Inadequate research-extension linkage was a limiting factor to adoption of this technology.

Oversowing, being a low-cost pasture development method, is likely to be widely used when and if the rightful settings are developed including its promotion through extension leaflets and demonstrations.

Conclusion

In this study, it has been shown that pasture development could be achieved through oversowing natural pastures with legumes. Oversown natural pastures were shown to be more productive, in terms of DM yield and LWG's, than untreated natural pastures. However, management of the oversown pasture and the forage legumes to be used varied with agro-ecological zones involved. Thus there was a need of extending this kind of study to cover the various zones of the country so as to develop this technology for each situation.

Acknowledgement

We are grateful for advice and assistance from various people in the Ministry of Agriculture and Livestock Development, the Tanzania Livestock Research Organisation (TALIRO), Sokoine University of Agriculture (SUA) at Morogoro and the International Livestock Centre for Africa (ILCA) in Addis Ababa, Ethiopia, notably Professor A.B. Lwoga and Dr. J.C. Tothill for their supervisory role to the senior author. Special thanks are due to Mr. J.J. Mbekelu for his diligence in the day-to-day supervision of the field work. We are indebted to the Royal Netherlands Government, Research Council, (Tanzania), World Food Programme (WFP) (Tanzania), International Development Research Centre (IDRC) (Ottawa, Canada), ILCA, TALIRO and SUA, for their support in this study. Last but not least, we wish to thank the PANESA and ARNAB Steering Committees for involving us to prepare this paper.

References

Anderson, G.D. and Naveh, Z. 1968. Promising pasture plants for northern Tanzania. E. Afr. Agric. For. J. 34:84-105.

Anon, 1983. The Livestock Policy of Tanzania. Ministry of Livestock Development 25 pp.

Anon, (1986). Annual Report. Livestock Production Systems (Tanzania) Project No. 3-P-84-007, TALIRO/IDRC.

Calo, L.L. 1976. Supplementary feeding. Proc. 3rd Scientific Conf. Tanzania Soc. of Anim. Prod. Mwanza, 1976:36-69.

Clatworthy, J.N. 1984. Effect of reinforcement of native grazing with Silverleaf desmodium (Desmodium uncinatum) on dry season performance of beef steers in Zimbabwe Trop. Grassl. 18:198-205.

French, M.H. 1957. Nutritional value of tropical grasses and fodders. Herb. Abstracts. 27:1-9.

Humphreys, L.R. 1981. Environmental adaptation of tropical pasture plants. MacMillan Publishers Ltd. London pp 261.

Jones, R.M. 1981. Studies on the population dynamics of Siratro: The fate of Siratro seeds following oversowing into short sub-tropical pastures. Tropical Grassl. 15(2):95-101.

Jones, R.M., Tothill, J.C. and Jones, R.C. 1984. Pastures and pasture management in the tropics and sub-tropics Tropical Grassl. Soc. of Australian Occassional Publication No.1:pp 74.

Keya, O., Olsen, F.J. and Holliday, R. 1971. Oversowing improved pasture legumes in natural grasslands of the medium altitudes of Western Kenya. East Afr. Agric. For. J. 37(2):148-155.

Kusekwa, M.L. 1982. Pasture improvement by oversowing. Proc. 9th Scientific Soc. Anim. Prod. Arusha, 1982.

Kusekwa, M.L. 1988. Establishment of pastures by oversowing. Paper presented at the 1st TALIRO Scientific Conf. Arusha, February 1988.

Kusekwa, M.L. and Lwoga, A.B. 1986. Establishment and early survival of nine pastures legumes oversown into natural pastures in northern Tanzania. In: I. Hague; S. Jutzi and P.J.H. Neate (eds): Potentials of forage legumes in farming systems of sub-saharan Africa. Proc. workshop held at ILCA, Addis Ababa, Ethiopia, 16-19th September, 1985:490-504.

Lane, I.R. and Lwoga, A.B. 1978. Pasture research at Morogoro. Proc. 6th Scientific Conf. Tanzania Soc. Anim. Prod. Tanga, 1978:134-146.

Lwoga, A.B. 1983. The potential of forage legumes in the exploitation of grazing resources in Tanzania. Paper presented at a symposium on the role of biology in the development and utilisation of natural resources in Tanzania, Dar-es-Salaam, 1983.

Mannetje, L't, 1984. Pasture development and animal production in Queensland since 1960. Tropical Grassl. 18(1):1-18.

Mannetje, L't., Jones, R.J. and Stobbs, T.H. 1976. Pasture evaluation by grazing experiment. In: N.H. Shaw and W.W. Bryan (eds). Tropical pasture research. principle and methods. CAB Bulletin 58:194-234.

Mwakatundu, A.G.K. 1977. A study on the nutritive value of East African pastures for ruminant livestock with special reference to mineral nutrition in grazing dairy cattle. Ph.D Thesis, University of Nairobi.

Naveh, 2. 1967. Introduction and selection of promising pasture plants for Arusha and Kilimanjaro regions of northern Tanzania. East Afr. Agric. for J. 32:96-102.

Northwood, P.J. and Macartney, J.C. 1969. Pasture establishment and renovation by direct seeding. East Afric. Agric. for J. 35:185-189.

Pratt, D.J.; Greenway, P.J. and Gwynne, M.D. 1966. A classification of East African Rangelands. J. Appl. Ecology 3:369-382.

Rukanda, N.A.K.; Lwoga, A.B. 1981. The role of legumes in improving natural grasslands for livestock production. Utilisation of low quality roughages in Africa Agric. University of Norway Aas. 185-189.

Stobbs, T.H. 1969. The effect of grazing management upon pasture productivity in Uganda. I. Stocking rate. Tropical Agric. (Trinidad) 46:187-194.

Tothill, J.C. Jones R.M. 1978. BOTANAL, a comprehensive sampling and computing procedure for estimating pasture yield and composition CSIRO. Division of Tropical Crops and Pastures. Tropical Agronomy. Technical Memorandum No.8.

Tothill, J.C. 1986. The role of legumes in farming systems of sub-saharan Africa. In: I. Hague, Jutzi and P.J.H. Neate (eds). Potential of forage legumes in farming systems of sub-saharan Africa. Proc. workshop held at ILCA, Addis Ababa, Ethiopia, 1619th September, 1985. 162-185.

Walker, B. 1969. Effects of nitrogen fertilizer and forage legumes on Cenchrus ciliaris pasture in Western Tanzania East Afr. Agric. For. J. 38:375-382.

Appendix Table 1: Rainfall during the study period (1979-1983) along with the 20 year rainfall means at Tengeru, Arusha, Tanzania.

Source: Meteorological facilities set-up experiment site (rainfall) figures for 1979-1983 at the Tengeru meteorological station (20 year rainfall means)

Appendix Figure 1. Monthly cummulative liveweight gains of heifers kg/head, 1981.

Appendix Figure 2. Monthly cummulative liveweight gains kg/head, 1982.

Appendix Figure 3. Monthly cummulative liveweight gains of heifers kg/head, 1983.

Appendix Figure 4. Monthly cummulative liveweight gains of heifers, kg/head and /ha. 1981.

Legend:

BR₁a = LWG/head at the high stocking rate
BR₁b = LWG/ha at the high stocking rate
BR₂a = LWG/head at the low stocking rate
BR₂b = LWG/ha at the low stocking rate

Appendix Figure 5. Monthly cummulative liveweight gains of heifers, kg/head and /ha 1982.

Legend:

BR₁a = LWG/head at the high stocking rate
BR₁b = LWG/ha at the high stocking rate
BR₂a = LWG/head at the low stocking rate
BR₂b = LWG/ha at the low stocking rate

Appendix Figure 6. Monthly cummulative liveweight gains of heifers, kg/head and /ha 1983.

Legend:

BR₁a = LWG/head at the high stocking rate
BR₁b = LWG/ha at the high stocking rate
BR₂a = LWG/head at the low stocking rate
BR₂b = LWG/ha at the low stocking rate