Effect of seedbed methods and time of harvest on the yield and nutritive value of some forage crops grown on Vertisol at Debre Zeit , Ethiopia

The influence of the improved drainage broadbed and furrow (BBF), as opposed to the traditional flat seedbed over the growing season on the dry matter yield and nutritive value of A vena sotiva, Vigna unguiculata, Lablab purpureas , Vida dasycarpa, Trifolium steudneri and Seshania sesban were studied on Vertisol. Up to 7 t/ha dry matter yield was recorded for Avena sotiva and Lablab purpureus when planted on Vertisols with improved drainage. Dry matter yield of forage crops also increased with advance in stage of maturity or subsequent harvests. On both improved and traditional flat seedbed methods, the chemical analysis of forage crops showed similar declines in crude protein content and in vitro dry matter digestibility (IVDMD) levels as the maturity of forage crops progressed.


Introduction
Integrated crop-livestock production is a long tradition in the Ethiopian highlands (above 1500 m altitude). Livestock provides draught power for tillage and transport and manure for crops while crop residues serve as important feed resources (Gry-SEELS 1988).
Vertisols (deep clay soils) in Ethiopia cover about 12million ha; 7.6 million ha are found in the highlands. Only about 2 million ha of Vertisols are under cultivation in the highlands. The remainder is not cropped because of waterlogging caused by high rainfall amounts (JUTZI 1986).
The traditional flat seedbed method is generally used on Vertisols by farmers. However, it cannot effectively overcome the problem of waterlogging (Berhanu 1985). Research at the International Livestock Centre for Africa (ILCA) on animal traction showed that the development ofan appropriate animal drawn implement (broadbed maker) for seedbed-shaping helped to overcome the problems of waterlogging. The broadbed maker forms the seedbed into broadbeds and furrows (BBF) which provides better drainage for cropping. On Vertisols and other soils in the highlands, native pasture and crop residue are the major feed resources. However, quality and quantity of these feedstuffs are insufficient to meet livestock requirements (Lulseged and Alemu 1984). Delayed harvest ofnative pasture for hay is also responsible for its low nutritive quality (Jutzi et al. 1987).
In order to alleviate these problems, ILCA's Highland Programme has focused its research on the evaluation of forage crops, especially legumes, for the highlands. In earlier studies, Avena saliva, Vigna unguiculata, Lahlah purpureas, Vida dasycarpa, Trifolium steudneri and Sesbania sesban have shown potential for the medium altitude (1500-2400 m asl) highlands of Ethiopia (Lulseged and Alemu 1984;Jutzi et al. 1987). These forage crops can provide high quality feed either as cut and carry during the rainy season or as hay for dry season feed when seasonal feed shortages become a serious impediment to meeting livestock requirements. These forage crops are high yielding (Abate 1984) compared with native pastures that have annual yields of only 3-4 t/ha (Jutzi et al. 1987) of dry matter, are poor in quality and result in low livestock performance in the dry season.
The role of forage legumes in increasing animal performance is well documented (Shaw 1961(Shaw , 1978Tothill 1974). However, the yield and quality of these forage crops are influenced by harvest time and drainage. In this study, field experiments were designed to examine the yield and quality trends of potential forage crops under the traditional and improved seedbed methods on Vertisol at different harvest times during the growing season.

Material and methods
Field experiments were conducted in the Ethiopian highlands, at ILCA's Debre Zeit Research Station (1800 m asl) from 1989 through 1990. The soil of the experimental site is Vertisol, which would normally present drainage problems in the high rainfall season. Rainfall occurs from mid June to mid September (main rainy season) and from Feb-ruary to April (small rainy season). Average annual rainfall is about 800 mm, with 70% usually falling in the main rainy season.
In the first year of the study, a randomised complete block design experiment with four replicates in a factorial arrangement oftwo seedbed methods using three forage legumes giving a total of 6 treatments each in a 3.6 x 20 m plot was established.
The flat seedbed represented the traditional system. The broadbed and furrow (BBF) was the improved drainage system intended to allow early planting on Vertisols and the removal of excess water from the seedbed.
In the second year, the experiment was sown as a split plot design. The two seedbed methods were used as main plots (each 18 m x 20 m) and he forage crops (one fodder oat and four legumes) were used as sub plots (each 3.6 m x 20 m plot) with six replicates ofeach. The forage crops studied and their seed rates are shown in Table 1. At sowing, 100 kg/ha di-ammonium phosphate fertiliser was applied uniformly to all treatments in both years.
Sampling was done using a quadrant frame of 1 m 2 in each sub plot which was repeated five times to provide yield estimation per net plot of 5 m 2 for each forage crop. Harvests were done randomly at 6, 12 and 18 weeks after plant emergence in the first year. Harvesting was carried out at 6,9,12,15 and 18 weeks after plant emergence in the second year in order to determine dry matter yield and Trifolium steudneri 13 nutritive value over the growing season. The dry matter data were subjected to analysis of variance (SAS); LSD and SED were used for mean comparisons.

Dry matter yield
Dry matter yields of the three forage crops in the 1989 crop season at different harvest times over the growing season are shown in Table 2.
There was no significant difference in dry matter yield between the flat and BBF seedbed methods at 6 weeks after emergence. However, at 12 and 18 weeks after emergence, there were significant differences in dry matter yield (P<0.05) between the forages grown on the two seedbed methods. Among the forage crops, dry matter yield differences were statistically significant at 6 weeks (P<0.001), 12 weeks (P<0.05) and 18 weeks (P<0.01) after plant emergence.
When harvested at 18 weeks, Lablab purpureas had the highest dry matter yield, followed by Vigna unguiculata. However, the performance of Vigna mguiculata at earlier harvests was the highest of the two forage crops.
The seedbed method by forage crop interaction on dry matter yield was significant at harvests of 6 weeks (P<0.05) and 12 weeks (P<0.05) after emergence. The interaction effect at 18 weeks after emergence was not significant.
Dry matter yields of the five forage crops in the 1990 cropping season at different harvest times over the growing season are given in Table 3. With the exception of the 18 week harvest time, there were no significant dry matter yield differences between the flat and BBF seedbed methods. There were highly significant differences (P<0.001) in dry matter yield among the forage crops at all harvest times. The interaction of seedbed method by forage crop was only significant (P<0.05) at 6 weeks after emergence.

Nutritional value of the forage crops
The chemical analysis of the forage crops for the 1989 season, sampled at different harvest times after plant emergence, is presented in Table 4. The chemical analysis of the forage crops in the 1990 season is shown in Table 5. In general, crude protein content and in vitro dry matter digestibility levels declined as the forage crops matured. The nutritional value of the forage crops at each harvest, using the flat and BBF seedbed methods, showed no statistically significant differences.
Thus, the values for the various quality parameters are presented as an average of the two seedbed methods.

Discussion
At Debre Zeit, an overall positive effect of the BBF was observed at the later stage of forage growth for the forage crops tested. This is reflected in the increased dry matter yields of forage crops over the traditional flat seedbed method. However, indi- vidual forage crop responses to seedbed method differed over the growing season (Table 2). The dry matter yields ofLablab porpureus (harvested at 6 and 12 weeks) and Vigna unguiculata (harvested at 12 and 18 weeks)appeared to be higher on the flat seedbed than on the BBF. In part, this can be attributed to the low rainfall (range 700-800 mm per annum), the low clay content (49%) of the soil (Kamara and Haque 1987)and the relatively high evaporative demands that cause varied drainage problems on different forage growth. These results are in contrast with those from higher altitudes(>24oo m asl) where clay content greater than 60% is recorded for Vertisols, rainfall exceeds 1000 mm per annum and temperatures are lower. These effects result in poor drainage on plant growth throughout the growing season (Juxzi 1983). Of the forage crops tested, Avena saliva, Lablab purpureas and Vigna unguiculata produced high dry matter yields (>6 t/ha) in both seasons.
The legumes are ofparticular importance for thenhigh forage quality (Table 4 and 5). Protein content was high at the earliest harvests but declined gradually as the legume matured. This result was ex-pected because older plants are more heavily lignified and normally produce poor quality forage.

Conclusion
The performance of these forage legumes in terms of yield and quality suggests they are crops of choice for integration into mixed croplivestock systems in the highlands. However, further studies are needed on how these forages are combined with food crops in land short systems in order to be used for livestock feed.
It is expected that if these forage legumes are successfully introduced into the highland cereal cropping system, they could have a significant impact on livestock production.