Teff (Eragrostis tef) plays a major role  in the livelihood of most  smallholder   farmers  in  Ethiopia.  It is  adapted  to diverse agro-ecological regions of Ethiopia and grows better under stress environments than do other cereals known world-wide (Hailu and Peat, 1996). The area covered by teff during the main season is about 1.91 million ha, which stands first in area coverage and accounts for 29% of the total cereal area in Ethiopia (CSA, 2010). However, its productivity is far lower than that of other major cereal crops growing in the country. Average national yield of teff in 2014/2015 production season ranged from 1281 to 1575 kg/ha (Sharma and Adera, 2016). Lower teff grain yield is mainly attributed due to inappropriate ploughing and weed control practices (Habtegebrial et al., 2007; Oicha et al., 2010). Other studies also indicated that weed is one of the key limiting factors for attaining higher teff yield (Mersie and Parker, 1983).

Tillage has been an important aspect of technological development in the evolution of agriculture, particularly in food production. Tillage systems are site specific and depend on crop, soil type and the climate (Rasmussen, 1999). Ploughing is one of the methods for soil tillage and weed control. Repeated ploughing makes the soil light and increase the available pore space for aeration in soil and root development. Due to this increase in available pore space, soil compaction is reduced and water holding capacity of soil increased (Bargali et al., 1993a, 2018, 2009) which enhanced the decomposition processes (Bargali et al., 1993b; Bargali, 1996), nutrient cycling (Bargali, 1995; Bargali et al., 1992; Bargali and Singh, 1997; Manral et al., 2020) and influences the crop productivity (Padalia et al., 2018). Some crops need very high tillage processes. Teff also needs high tillage frequencies as compared to other cereal crops in Ethiopia.  It also requires firm, and level seedbed, free from clods and stumps. With respect to teff cultivation on vertisols, several ploughings are necessary, occasionally as much as 12 times, relative to nitosols (Deckers et al., 1998). According to Kenea et al. (2001) the tillage frequency for teff in Ethiopia ranges from three to twelve times. Though some research results indicated teff grain yield  increased   with   increasing number  of  ploughings (IAR, 1998; Melese, 2007).

Generally, farmers practice conventional tillage to control weeds, conserve moisture, and increase soil warming (Oicha et al., 2010). Other researchers (Mulumba and Lal, 2008; Rachman et al., 2003) noted that it is essential to select a tillage practice that sustains and favors successful growth of agricultural crops in a given environmental condition. An experiment conducted at DebreZeit Research Center in Ethiopia indicated that hand weeding once at early tillering or twice at early tillering and stem elongation stage depending on the degree of weed infestation of the crop was profitable (Seyfu, 1993). There is inadequate information on the effect of ploughing frequency and weed control practices on teff growth and yield in the low land areas of Gamo Zone. Therefore, the objective of this study was to evaluate the effects of tillage and weeding frequency on growth, yield and yield components of teff varieties.

Description of the study area

The experiment was conducted at Fura kebele in Mirab Abaya woreda, Gamo Zone in Southern Ethiopia on a farmer’s field during 2019 short cropping season. The place is located in between 6°11′0′′ - 6°13′0′′N and 37°37′0′′ - 37°45′0′′E longitude and latitude, respectively and at an elevation of 1193 masl with the average annual rain fall of 610 mm and average minimum and maximum air temperatures of 19 and 34°C, respectively. The area is located at 30 km away from Arba Minch city in northern direction. Soil properties of the study site are described in Table 1.

Treatments, design and experimental procedures

The experiment consisted of three levels of tillage frequencies (one pass, four passes and six passes), three levels of weeding frequency (no weeding, one time weeding at tillering and two times weeding at tillering and stem elongation stages of the crop) and two varieties of teff (local and improved variety namely Quncho or DZ-Cr-387-RIL 355). Quncho was developed and released by Debrzeit Agricultural Research Centre in 2006, it is a high yielding variety (1.8-2.6 t/ha  on-farm),   early  maturing  (86-151 days),  and  white- seeded cultivar adapted to a wide range of altitudes (MoA, 2010). The treatments were laid out in split-split plot design using tillage as a main plot factor, weeding as subplot factor and variety as sub-subplot factor with three replications. Those plots, which received six plowing frequencies, were plowed first at late January, 2020; second up to fifth plows were done within fifteen days interval and the sixth at sowing. The plots, which received four plowing frequencies, were plowed first, at the late February, 2020; second up to third plows were done within fifteen days interval and the fourth at sowing. The plots, which received one plowing, were plowed only at sowing.

After preparing the land as per treatments required, each main plot divided into three sub plots (1.2 m × 4 m) and finally each sub plots divided into two sub sub-plots (1.2 m × 2 m) with net plot size of 1.2 m × 2.0 m².  A plot size of 1.2 m × 2 m (2.4 m²) with 10 cm row spacing and a total of 10 rows were used. The net harvestable area was 1.0 m × 1.8 m and adjacent plots and blocks were spaced 0.5 and 1 m apart, respectively. All agronomic practices were conducted as per recommendations except the treatments.

Data collection

Data were collected on weed species composition and density as well on crop parameters including days to 50% panicle emergence, days to 90% physiological maturity, total tiller number/m², plant height (cm), number of effective tillers/plant, panicle length (cm), total dry biomass yield (kg/ha), grain yield (kg/ha), harvest index (%) and thousand seed weight (g).

Statistical analysis

Data were subjected to analysis of variance (ANOVA) using General Linear Model (GLM) procedures of SAS 9.0 (SAS, 2004). Differences among treatment means were compared using Least Significant Differences test at 5% level of probability.

Economic analysis

The economic analysis was done to determine the economic feasibility of the treatments following the procedure developed by CIMMYT (1988). Costs of tillage, weeding and teff seed were taken as variable costs during the experiment and prevailing market price of the crop was estimated at the time crop harvest as return. Seed yield was also adjusted down by 10%.

Weed species composition

Weed species recorded during the study period were grasses and broad-leaved species of herbs. The dominant weeds were grasses in the experimental site (Table 2).

Effect of tillage frequency on weed density

Weed density was significantly (P< 0.05) influenced only by the main effect of tillage frequency (Table 3). Weed density was significantly reduced with increase in the number of tillage passes. 52 and 50% reduction in weed density was observed due to six and four tillage passes, respectively, compared to the value obtained single pass plow. Accordingly, maximum value of weed density (9 plants/m²) was recorded due to one pass tillage. This result is in agreement with report of ICARDA (1984) which shows that repeated tillage compared to no-till reduced weed infestation by 50%.

Days to 50% panicle emergence

This parameter was significantly influenced only by the main effects of ploughing frequency and variety (Table 4). Statistically similar and longer durations of days to 50% panicle emergence were recorded due to fourth plowing and six times compared with single pass plowing. The increment in duration of panicle emergence with increase in plowing frequency might be associated with easy and efficient utilization of soil nutrients and moisture from fine seedbed which favored prolonged time for vegetative growth with delayed heading of the crop. Regarding varietal effect, local variety took statistically longer duration to (36.8 days) 50% panicle emergence compared with improved variety (Quncho). This variation might be due to genetic difference between the two varieties as reported by Seyfu (1993).

Days to maturity and growth parameters 

Days to maturity, total number of tillers/m², number of effective tillers/plant and plant height were significantly affected by the interaction effects of plowing frequency, weeding frequency and variety (Table 5).

Days to 90% physiological maturity

The maximum number of days to maturity (69.67 days) was observed for treatment combination of ploughing six times with two times hand weeding at tillering and stem elongation stage for the local variety, while the lowest value (65 days) was recorded for ploughing once combined with un-weeded control for the improved variety. The prolonged period for maturity with increased ploughing frequency and weed free environment or minimized weed population  might  be  due  to  prolonged period of vegetative growth at the expense of reproductive development as a result of probably improved nutrient and moisture supply to the plants. Varietal difference in maturity period could be attributed to differences in the genetic makeup. This result was in agreement with the findings of Haftamu et al. (2009) who reported that teff requires 90 to 130 days for growth depending on variety and altitude. In contrast, Evert et al. (2008) reported that little weed competition due to suitable agronomic management practice results in fast growth and earlier maturity of crops.

Total tiller number

The highest total tiller number (182.67) was recorded for the combination of six plows with two times hand weeding for the improved teff variety, followed by the same treatment combination for the local teff type (177.67) (Table 5), while one plow with un-weeded control resulted in lower values for both local (95.67) and improved (122.30) varieties. This might be due to more availability of soil nutrients and other environmental inputs for the crop due to repeated tillage and in the absence of weed competition. The result was in line with the findings of Aleligne (1988) who reported that fine seedbed preparation is to get weed free environment and good crop germination. Similarly, Tenaw (2010) has reported that repeated tillage created conducive environment for better plant establishment, growth and production because of reduced weed competition and possibly better root growth.

Plant height

Maximum   plant   height   (114.3 cm)   was  recorded  for minimum tillage combined with un-weeded plot for the local variety, while the lowest value (89.3 cm) was recorded for the same variety with one plow pass combined with weeding once at tillering stage (Table 5).

This may be due to fast vertical growth of the crop to successfully compete with weeds for light and space.  The result of the present study was in agreement with the findings of Bekele et al. (2018) who reported that increased plant height with the weedy plot might be due to the effect of severe competition among plants which make them elongated in search of light.

Number of fertile tillers

The highest number of effective tillers (7.3) was recorded for six plows with two times weeding for the improved variety, followed by (7) four times plowing combined with two times weeding for the local variety, whereas the lowest value (2.6) was recorded for minimum tillage (one plow pass) with un-weeded treatments for the local teff type (Table 5). Number of fertile tillers increased with increasing tillage and weeding frequency, probably because of increased nutrient and moisture supply and more  light  to  teff  plants  as  a  result  of  reduced  inter-specific competition between the crop and weed species. The current result was in agreement with the findings of Gebreyesus (2014) who reported that number of effective tillers significantly increased as tillage frequency increased from minimum tillage up to four times plowing.

Thousand seed weight and harvest index

Thousand seed weight was significantly influenced only by the main effect of weeding frequency whereas harvest index insignificantly responded to main and interaction effects of the treatments (Table 6). Treatments in which weeds were removed resulted in higher and statistically similar values of thousand seed weight compared with control (un-weeded treatment). This might be attributed to better availability of growth resources for the crop in weeded treatment which results in better assimilate production and translocation to sink (seed).

Yield and yield components

Panicle length, above ground biomass and grain yield were significantly influenced by  the  interaction  effects of ploughing frequency, weeding frequency and variety (Table 7).

Panicle length

The longest panicle (39.0 cm) due to four times ploughing and twice weeding in local variety whereas the shortest panicle (27.30 cm) was observed due to combined effect of one pass ploughing, un-weeded (control) treatment and local variety (Table 7). The increase in panicle length with repeated tillage and frequent weeding might be attributed to better availability of growth resources for the crop resulting from repeated tillage and weeding which is in agreement with the findings of Reicosky and Allmaras (2003) and the genetic nature of varieties.

Biomass yield

The highest biomass yield (6759.2 kg/ha) was obtained from treatment combination of six times plowing with two times hand weeding at tillering and stem elongation stage for the improved teff variety, while the lowest value was 3888.8 kg/ha followed by 3981.4 kg/ha recorded for one-time plow and un-weeded plot for the improved variety (Quncho) indicating that 74% variation in above ground biomass due to treatment interaction (Table 7). The increase in biomass yield with increasing tillage and weeding  frequency   could   be   attributed   to  increased availability of nutrients and soil moisture for the crop as a result of reduced competition by weeds as reported by Alemu et al. (2016). On the other hand, the lowest biomass yield (3888.9 kg/ha) was recorded from combined effect of one plow and un-weeded plots in improved variety implying the sensitivity of modern varieties to poor agronomic practices.

Grain yield

The highest grain yield (1192.53 kg/ha) was recorded for six times (conventional) plowing combined with twice hand weeding for the improved variety (Quncho) followed by the value (1135.17 kg/ha) obtained from four times plowing with weeding twice at tillering and stem elongation stage for the local variety, which produced the lowest grain yield (775.93 kg/ha) when un-weeded and grown with one plow pass (Table 7). The increase in grain yield with frequent tillage and weeding might be due to enhanced uptake of essential nutrients by teff plants as a result of reduced competition of weeds as reported by Alemayehu et al. (2008). Variation in grain yield (54%) was recorded due to interaction effects of ploughing frequency, weeding frequency and varieties.  The current result also evidenced by the findings of Gebreyesus (2014) and Alemayehu et al. (2008) who reported higher grain yield under frequent ploughing and weeding, respectively. Grain yield showed highly significant and positive linear relationship with total tiller number (r= 0.72**),  effective   number   of   tillers (r= 0.59**),  panicle length (r= 0.57**), thousand seed weight (r= 0.36**), days to physiological maturity (r= 0.35**) and above ground biomass (r= 0.66**) implying that improvement in these parameters is also associated with better grain yield.  

Economic analysis

The results of economic analysis indicated that the highest  net  benefit  (Table   8)   was  obtained  from  the treatment of one plow pass combined with twice hand weeding at tillering and stem elongation stage on local variety due to its lower cost of production (about 48% lower) but its marginal rate of return (MRR) of 861% was lower than that of four-time plows combined with twice weeding for variety Quncho (5800%). According to the current result, plowing the soil more than once is important for teff production and productivity.

Four times tillage passes showed better yield and yield components as compared to the others. Similarly, twice hand weeding showed significantly higher yield of teff. improved variety (Quncho) resulted in a better yield compared to the local variety for most treatment combinations. In general, four times oxen plowing combined with twice hand weeding (at tillering and stem elongation stage) gave the highest MRR (5800%) and NB (ETB 27503.4 ha^-1) for variety Quncho as compared to other treatment combinations. Therefore, four ploughing, twice hand weeding (at tillering and stem elongation) and Quncho variety should be recommended for higher productivity and profitability of teff in Mirab Abaya, Southern Ethiopia.

Hence, this practice was found to be the best option that could be recommended for teff production by using improved variety (Quncho) in the study area and in areas with similar agro-ecology. However, it is advisable to undertake further research across different soil types, seasons and locations to draw a more comprehensive recommendation.

The authors would like to thank Arba Minch University for funding this research. They also thank Gamo Zone Environment Protection, Forest and Climate Change Controlling office for their help during study leave to the first author. 

The authors have not declared any conflict of interests.