1. Introduction
Maize [Zea mays L.] is a cereal crop that is cultivated world-wide in a wide range of agro ecological zones. It was introduced into Africa in the 1500s and has since become one of Africa’s dominant food crops. It is the most important cereal crop of the world, grown in irrigated and rain-fed areas [1] . Maize is considered a very significant cereal crop and probably has the greatest yield potential among crops in sub-Saharan Africa (SSA) as it occupies more than 50% of aggregate land given entirely to crop production [2] . Due to the ease of processing, high yield potential, and low cost of production, maize is considered very significant in ameliorating food insecurity in SSA [3] . In industrialized countries, maize is largely used as livestock feed and as a raw material for industrial product. According to [4] , maize is the second most important commodity crop in Ghana after cocoa accounting for more than 45% of the agricultural cash income among smallholder farmers in the country [5] . It is the most widely consumed staple food with increasing production since 1965 [6] and is a major source of food for both urban and rural dwellers in Ghana. It forms part of everyday diet and also serves as an important raw material for animal feed. Maize consumption accounts for more than one-quarter of calories consumed, about double that of the second crop, cassava [7] . Maize consumption is projected to increase due to population growth. Increased rate of population growth coupled with urbanization, poultry keeping and fish farming have contributed to increased demand for maize in Ghana. The poultry industry’s demand for maize, used as feed, was estimated to have grown by 10 percent annually between 2000 and 2009. About three-quarters of maize consumption is from own production, suggesting maize has limited appeal as a cash crop [8] . This is set to change as Ghana’s Planting for Food and Jobs (PFJ) initiative, launched in 2017, prioritizes maize seed and fertilizer distribution and encourages market participation by smallholders [9] .
Despite the importance of maize, there are several biotic and abiotic constraints limiting its production. These include the parasitic weed; Striga, causing food scarcity faced by 100 million people, thereby increasing economic challenges that amount to more than 10 billion USD losses yearly [10] [11] . In the Guinea Savannah zone of Ghana, a major threat to maize production is root parasitic weed of the genus Striga in the family Orobanchaceae, [12] . Itis endemic in the zone causing poor grain yield to poor resource maize farmers, [13] . The threat is aggravated by climate change with concomitant negative effects on food security and the general livelihood of the people in this part of the country. Striga infestation flourishes in conditions characterized by low soil fertility, and continuous monocropping with cereals. Crops such as soybean act as false hosts and trap crops. Trap crops stimulate the Striga seeds to germinate but the Striga seedling cannot successfully attach its roots to the roots of the trap crops in order to feed and hence Striga dies in the process. Crop losses due to Striga cause yield reduction of up to 100%, excluding the high effect on the agricultural product [14] [15] . S. hermonthica, generally known as purple witch weed is a challenging weed to crops in SSA [16] . It is a hemiparasite of grasses belonging to the family Orobanchaceae. This species of Striga causes a destructive harvest reduction annually and attacks crops, such as sorghum, sugarcane, finger millet, pearl millet, rice, cowpea, and maize, causing a decimating effect on the yield of crops [17] . This parasitic weed depends solely on its host plant for nutrients and growth, which results in substantial damage to crops, such as chlorosis, wilted silk, thin stalk, reduced height, and total loss of crops in farmlands with high invasion [18] [19] . The invasion of this parasite is prominent in places with characteristic poor soils and intense cultivation with poor management practices [16] .
Several methods, such as the traditional, chemical, and biological methods, and host resistance have been used to combat this parasitic weed. Traditional practices including hand pulling, crop rotation, intercropping, trap or catch crop planting, push-pull technology, and soil fertility improvement (nitrogen fertilization) are used. Chemical management practices, for example, weedicide (imazapyr), genetic control, resistant crop varieties, suicidal germination, biological control agents, such as fungi and bacteria, and certain insects have been used [20] [21] [22] . None of these methods have been able to eradicate the incidence of S. hermonthica alone, however, the incorporation of two or more methods have proven effective in eradicating it completely. The adverse effect of chemical compounds on soils has led to increased use of microbial inoculants, which are advantageous to plant development by enhancing its secretion of plant growth hormones, increasing nutrient uptake, and reducing plant pathogens [23] . In their efforts to help farmers manage Striga infestation on their farms, the Crops Research Institute (CRI) and the Savannah Agricultural Research Institute (SARI), both of the Council for Scientific and Industrial Research of Ghana have developed a range of maize varieties purported to be tolerant/resistant to S. hermonthica. SARI also released the soybean variety “Afayak” (i.e. TGX 1845-5E) targeted for use as Striga trap crop that could accelerate Striga germination with the long-term effect of depletion of Striga seed banks in Striga endemic fields [24] . However, there is no report on the adaptive and relative performance of these varieties in integration with the trap crop (Afayak) in the Guinea and Sudan savannah in Ghana, for which they have been developed. The purpose of this study was to evaluate the effects of integrating soybean (Afayak)-maize on Striga management in the Guinea Savannah zone of Ghana. Intercropping Striga tolerant maize with soybean will give farmers better and cheaper way to control Striga hermonthica. In Ghana intercropping legumes such as soybean with cereals is a common practice by farmers and the practice has been reported to reduce Striga seed bank in Striga endemic locations of Northern Ghana [25] . The objective of this study was therefore to evaluate the effectiveness of maize-soybean integration in managing the effects of Striga hermonthica in the study area. Specifically, the objectives were to:
1) Assess Striga tolerance of the four maize varieties in terms of maize-soybeanyield and yield components.
2) Identify the most effective maize variety in combination with the soybean (Afayak) for the reduction of Striga seed bank.
2. Materials and Methods
2.1. Experimental Site and Design
The trials were conducted at Gore near Binaba, in Zebila of Bawku West District of the Upper East Region of Ghana on a farmer’s farm in the 2015 and 2016cropping seasons. This area was chosen because it is heavenly infested by Striga hermonhtica. The experimental site was located on latitude 10˚48"N and 0˚28"W and longitude 0˚33'1"W (IFDC personal communication 2015). The vegetation at the site is savannah grassland which is characterized by shrubs and few scattered trees. The climate is warm, semi-arid with a total average annual rainfall of about 1100 mm [26] . There is a short rainy season followed by a long dry season which is between October and April. The climatic condition is characterized by two air masses, the North East trade winds usually dry and the South West Monsoon winds. The trials were 2 × 4 factorial laid out in Randomized Complete Block Design (RCBD) with three replications. The eight treatments consisted of four drought and Striga tolerant maize (Aburohemaa, Omankwa, Bihilifa, and Wang dataa) varieties planted with and without the recommended soybean variety (Afayak) (Table 1). A replication was made up of eight plots separated by 0.5 m with the alley between blocks and replications being 1.0 m and 2.0 m respectively. Plot-size of 10 m × 10 m was used with total experimental area of 2780.50 m2 (67.0 m × 41.5 m).
2.2. Management Practices
During the major cropping seasons (2015 and 2016), the experimental area was ploughed, harrowed and ridged (in July) using bullock-drawn implements. Lining and pegging were done to establish the plots for the treatments. The ridges
Table 1. Entries for researcher-managed trial in 2015 and 2016.
were separated by a distance of 1 m. At planting, three seeds per hill were sown at an intra-row spacing of 50 cm inter-row spacing of 75 cm. The three seeds per hill were later thinned to one plant per stand. Weed control was done by hoeing twice at 2 WAP and 5 WAP which coincided with fertilization. Fertilizer grade NPK 15-15-15 was applied at 15 kg N/ha, 15 kg P2O5/ha and 15 kg K2O/ha at 2 weeks after planting (WAP) and 5 WAP; (half N was used for top dressing). Fertilizer application targeted maize hills only.
2.3. Data Collection
2.3.1. Soil Analyses
The soil characteristics were determined in order to know nutrients status and Striga load of the experimental site before application of the fertilizers. Three composite soil samples were taken for determination of physical and chemical properties. At the beginning of the experiment (in 2015), 15 samples were randomly collected by using an auger and composited. Then, soil samples were also taken from each treatment at harvesting (in 2016). The samples were air dried, crushed with mortar and sieved to pass through 2 mm mesh. The characteristics analyzed for included; Soil pH, organic matter, total nitrogen, exchangeable calcium, magnesium, potassium, sodium and effective cation exchange capacity, and bray No. 2 extractable phosphorus and potassium. The air-dried soil samples were ground at the laboratory and sieved through a 2 mm sieve. Soil pH was determined using a glass electrode (pH meter) in a soil ratio of 1:2.5 as reported by [27] and [28] . Soil organic matter was determined by the wet combustion method [29] . Percentage total nitrogen was determined by the micro Kjeldahl-technique [27] . The available phosphorus was extracted by the Bray method and determined colorimetrically [30] . Potassium was determined by flame emission photometry [27] . The exchangeable cations calcium, magnesium, potassium and sodium were determined as recommended by [27] using EDTA Titration after extraction with 0.1N Ammonium Acetate at pH 7. Effective Cation Exchange Capacity (ECEC) was calculated as the sum of the exchangeable bases and exchangeable acidity [27] .
2.3.2. Plant Height
Data were collected on plant height at two weeks interval from 5 WAP to 13 WAP, on five tagged plants in the middle part of each plot. Maize height was determined by measuring from the ground level of the stem to the last emerged leaf using a tape measure.
2.3.3. Number of Leaves of Maize Plants
Maize and soybean leaves were counted on the five (5) tagged plants in each plot. Leaves were counted on each plant starting from the first leaf from the base of the plant to last emerged leaf. Maize leaves were counted 7 and 9 WAP.
2.3.4. Leaf Area Index (LAI)
Leaf area index of maize was determined once in the 9 WAP and was taken by measuring the length and width of three (3) leaves from the lower part of the plant, in the middle and at the top on each of the 5 tagged plants in each plot. Each leaf area index was estimated using this formula; Leaf area index = 0.75 × LL × LW; where 0.75 is the constant for maize, LL = leaf length, LW = leaf width.
2.3.5. Length of Cob, Height of Cob Attachment and Cob Weight
Height of cob attachment was done by measuring from the base of the plant to the point where the cob attaches to the maize stalk using tape measure. Similarly, a tape measure was used to measure the length of cobs from the point at which the cob attaches to plant to the tip of the cob before the maize was then harvested. Five (5) cobs were selected from the five (5) tagged plants in each treatment and weighed and their averages recorded. An electronic scale was used to determine the cob weight at harvest.
2.3.6. Striga Shoot Count
Striga seedlings were counted as they emerged in every two (2) weeks interval, starting from the 7 to 16 WAP. At each Striga count, the plants of the parasite were uprooted.
2.3.7. Harvesting
Maize grain and stover were harvested at maturity from a net area of each treatment demarcated after leaving out two rows on each side of the plot and the first two lines of maize/soybean plants on each row to minimize the edge effect. The entire plants on the plots were harvested by cutting at the ground level. Maize cobs were manually separated from the stover, sun-dried, and packed in sacks before threshing. Weights of maize grain yield per plot were taken using a weighing scale. Striga count and leave count were transformed using the square root (
) transformation.
2.3.8. Formulas Used
(1)
(2)
(3)
(4)
Postharvest soil samples were also taken for analysis of Striga seed bank as a result of soybean-maize intercropping systems.
2.3.9. Analyses of Data
The data were statistically analyzed using the Analysis of variance (ANOVA) for randomized complete block design (RCBD) and the Least Significant Difference (LSD) was used for mean separation (P ≤ 0.05) following the procedure of Steel and Torrie (1980). Count data were transformed using
before subjecting them to analysis of variance.
3. Results and Discussion
3.1. Plant Height of Maize
Plant height of the four maize varieties with or without integration with soybean was not significantly (p > 0.05) different for both 2015 and 2016 cropping seasons at the measured growth stages on all sampling days (Table 2). Optimum plant height was attained at 9 WAP, but slight increases were generally observed up to 13 WAP during which the results of plant height, ranged from 145.9 cm (Omankwah) to 161.9 cm (Bihilifa), suggesting that the expected and biotic stress imposed by S. hermonthica did not suppress maize height. The maize varieties were bred for drought and Striga tolerance, and this could be the reason for the observed expression of this trait. It could also be the genetic makeup of the varieties used for the study. This observation is consistent with reports by [31] who did not find any significant difference in plant height of maize under mono-cropping and intercropping with sugar-beet and groundnuts. Thus, the treatment differences were not significant in affecting maize plant height in the current study implying that presence of Striga in the experimental plots did not affect height of maize. Although, Striga reduces cell elongation as it takes photosynthesis away from the maize leading to shorter maize internodes and stunted growth, yet these symptoms were not observed on the plants in this study. Thus, Striga did not have negative effect on the crops. This agrees with [32] [33] [34] , who reported that intercropping sorghum and maize with legume crop especially Desmodium spp, significantly enhanced both plant height and grain yield in maize.
Table 2. Plant height of maize during the 2015 and 2016 cropping seasons.
3.2. Plant Population of Maize at 9 WAP and at Harvest
Similar to plant height, plant population of the four maize varieties with or without integration with soybean was not significantly (p > 0.05) different for both 2015 and 2016 cropping seasons at the measured growth stages on all sampling days (Table 3). Results of plant population per unit area at harvest showed no significant effect by treatments. It was observed from field inspection that in plant stand of maize, all the maize varieties used germinated well. Also, it means birds and rodents did not remove the seeds/seedlings at the initial stages. This result was not in agreement with the findings of [33] who reported that both plan population density and variety showed significant difference in final plant population of maize. The differences in the two studies might be as a result of genetic materials used or might also be that, some of the seeds used in the study of [35] were not viable. The eight treatments did not show significant (p < 0.05) difference in leave area index (LAI). LAI ranged from 5.0 cm to 5.6 cm. In a similar study, [31] also did not find any significant differences in LAI between maize monocrop and maize intercropped with sugar beans or groundnuts. These findings are consistent with results of [36] who did not find any significant differences in LAI between sole maize and maize intercropped with cowpea. The leaf area (LA) describes the size of the assimilatory apparatus of a plant stand and is the main factor that determines the rate of dry matter production in a closed stand. It also reflects differences in productive efficiency between crop varieties [37] . The non-significant difference of leaf area observed in the present study could be due to the genetic makeup of the varieties, environmental conditions, and cropping systems used. This is in line with [38] that these four maize varieties are resistant/ tolerant to S. hermonthica infestation.
Table 3. Plant population of maize at 9 WAP and at harvest and LAI 9 WAP during the 2015 and 2016 cropping season.
3.3. Striga Count
Statistically, no difference (p > 0.05) was observed on the impact of the eight treatments on Striga count at 7 WAP. Except for the low Striga emergence of 7 plants/plot under Bihilifa/Soybean, similar Striga count was observed across the cropping systems. Wang dataa supported the highest Striga emergence of 52 plants/plot and the intercrop Bihilifa/Soybean the lowest. Striga emergence was pronounced in sole Bihilifa and Omankwa/Soybean in the 13 WAP (Table 4). The high Striga numbers in some of the treatments might be attributed to high initial Striga seed bank at the site. This might also be due to variation in soil fertility where some plots might have low soil fertility. Striga thrives well in less fertile soils as supported by [39] who reported that one of the witch weed most contributing factors for development is low soil fertility and cropping systems in SSA with no external inputs. Because Striga is an obligate parasite, interactions between Striga and its host plays a crucial role in survival of the parasite, if this interaction was disrupted, it might be a beneficial approach for integrated management of this parasite. Differences in production of Striga stimulants are known to occur between crop cultivars [40] , and that may be the cause for reduced Striga emergence in some of the treatments in the current study. The low number of Striga plants in some of the treatments could be due to their ability to show some levels of resistance to the parasitic weed, which reduced the extent of severity of Striga infestation. This is supported by a baseline study carried out by [41] looking at the extent of Striga infestation on maize grown in Western and Nyanza provinces. It was also observed that, some maize plots intercropped with soybean resulted in high numbers of Striga emergence count than some of the maize plots that were sole cropped. This was probably because of the soybean variety (Afayak) which has the ability to cause germination of Striga seeds but do
Table 4. Striga count at 7 and 13 WAP for the 2015 and 2016 cropping season.
not support it subsequent growth and development. This present study is not in line with [42] who reported that the maize varieties grown in Ghana under intercropping supported fewer Striga infestation compared to those grown in sole cropping.
At 9 and 16 WAP, Striga count varied with treatment (p < 0.05) such that Omankwa/Soybean in 9 WAP recorded the highest Striga emergence but similar to Wang dataa/Soybean (Figure 1). Maize monocrops and intercrops of Aburohemaa/Soybean and Bihilifa/Soybean recorded similar emergence of Striga seedlings. It was not clear at this stage which cropping system was outstanding for enhanced or reduced Striga seedling emergence. Wang dataa/Soybean slightly had highest percentage (5.3%) Striga emergence in the 16 WAP, but not significantly different from Aburohemaa/Soybean (4.9%) and Bihilifa/Soybean (4.8%), sole Aburohemaa (4.5%) and sole Bihilifa (4.5%) which all supported Striga emergence. Sole Wang dataa and Omankwa/Soybean equally supported Striga emergence with Sole Omankwa recording the lowest Striga emergence in the 16 WAP.
It was interesting to note that, the Striga with higher numbers in germination or emergence did not show any negative effect on the crops which might be due to maize resistance/tolerance level to the witch weed. On the other hand, some plots with sole maize exhibited resistance to S. hermonthica by supporting the lowest number of Striga plants germination unlike some of the maize-soybean intercropped with greatest number of Striga emergence. Possible reason for this could be due to Striga seeds which did not germinate because of absence of chemical stimulant. Also, the Striga seeds will not germinate unless they have been conditioned, that is., are no longer dormant and are exposed to the right environmental conditions for germination. This result agrees with [43] who reported that nitrogen reduced the severity of S. hermonthica. It also agrees with the findings of [44] who reported that maize resistance to the Striga is the eventual expression of a series of interactive events between the parasite and its hosts. Similarly, [45] reported that resistant crop genotypes support significantly fewer
Figure 1. Striga count at 9 WAP during the 2015 and 2016 cropping season.
emerged of S. hermonthica plants. This study confirms the findings of [44] who reported that resistant varieties effectively reduce the Striga count with or without other options, indicating that host plant resistance alone could be used in situations where integration of all options is impossible.
3.4. Striga Biomass Production
Striga biomass production at 7, 13, and 16 WAP was not significantly different (P > 0.05) among treatments (Table 5). The highest biomass produced was recorded in the 13 WAP during which sole Bihilifa produced (422 kg/ha), Omankwa/soybean (360 kg/ha), and Aburohemaa (285 kg/ha). The lowest biomass (17.0 g/ha) was produced in the 16 WAP by Omankwa (Table 5). Greater number of Striga biomass was observed in sole Wang dataa, Wang data intercropped with soybean, sole Bihilifa, and sole Omankwa. The greater Striga biomass might be due to initial Striga seed bank variations at the site and the more the seed bank the more seeds will germinate with suitable hosts, hence, translating to greater Striga biomass. The greater Striga biomass might also be due to high crop density with high host and soybean root surface area. According to [46] , the level of Striga biomass on a host influences host productivity, but added that the relationship is non-linear; that is a point is reached where host grain production is independent of parasite biomass. The greater Striga biomass in some of the plots could also be due to variation in soil fertility status of the site as Striga thrives well in soils with poor fertility. The reduction in Striga biomass in some of the treatments could be due to reduction in number of Striga plants emerged which might also be due to the differences in the initial Striga seed bank, and soil fertility. Here again, though Striga emerged in all the plots, yet treatments did
Table 5. Striga biomass production at 7, 13, 16 WAP during the 2015 and 2016 cropping season.
not have any Striga symptoms. It implied that these treatments had more tolerance level to S. hermonthica infestation. This result agrees with [46] who observed that highest Striga infestation did not necessarily translate into yield reduction. Generally, the results indicated reduction of Striga biomass in intercrop maize and increase for the sole maize which might be due to the shading effects in the intercrop. This observation is in line with [47] who reported that striga biomass reduction may be due to shading effects from the maize-soybean intercropped plots. In a similar work, [48] observed that intercropping reduced the Striga biomass by 25% - 65% and 10% - 80% during the first and second season respectively. In fresh weight of Striga, sole Wang Dataa had 46 g/ha, followed by Omankwa/soybean (36 g/ha) similar to Aburohemaa/soybean (36 g/ha) whilst sole Aburohemaa had the least (12 g/ha). The dry weight of Striga followed a similar trend in the treatments in fresh weight of Striga. It still means that those treatments were highly tolerant to Striga negative effects because crop growth was not affected.
3.5. Percent Reduction in Striga Seed Bank
Percent reductions in soil Striga seed bank varied across treatments (p < 0.05), ranging from 30.3% (Aburohemaa) to 52.8% (sole Omankwa) (Figure 2). Though, S. hermonthica seed bank was high at the initial stage, yet at the postharvest S. hermonthica seed bank reduced between 5% - 26% across all the treatments. The high seed bank load before planting or harvest might be due to the level of S. hermonthica plants that flowered and produced seeds previously. This is in line with [49] who reported that Striga seed bank is determined by the level of Striga plants that flower and produce seeds, coupled with lack of suicidal germination. The high seed bank load may also be favoured by mono cropping because mono cropping of cereal hosts with little or no specific measures against Striga would lead to huge amounts of seeds accumulating in the seed bank.
The decreased number of S. hermonthica seed bank in the maize/soybean intercrop may be attributed to the suicidal germination caused by the germination
Figure 2. Percent reduction in Striga seed bank after harvest.
stimulant produced by the soybean (Afayak) roots. This is in line with report by [50] that the use of trap crop such as soybean triggers suicidal germination of Striga and therefore reduces the Striga hermonthica seed bank in the soil when intercropped with maize. In addition to being a trap crop, soybean provides shade which smothers the Striga thereby reducing its vigour. This result indicates that these Striga tolerant maize varieties when planted sole can help reduce Striga seed bank in Striga endemic areas. The reduced S. hermonthica seed bank in the soil in both cropping systems during the cropping seasons means a reduced potential for overall flower and capsule production and, consequently, a reduced capacity of increasing the S. hermonthica seed bank in the soil. The Striga infestation did not affect yield significantly. This result indicates that these Striga tolerant maize varieties when planted sole can help reduce Striga seed bank in Striga endemic areas in the near future, as more will germinate whose growth and development are not supported. The reduced S. hermonthica seed bank in the soil in both cropping systems means a reduced potential for overall flower and capsule production and, consequently, a reduced capacity of increasing the S. hermonthica seed bank in the soil. An effective management approach for Striga should therefore aim, among other things, to reduce and eventually deplete the soil seed bank.
3.6. Cob Weight, Grain Yield, Harvest Index
These parameters were not statistically affected by treatments. This could be so as a result of inherent ability of the four maize varieties in performance in terms of yield and yield components. It could also be as a result of the low fertilizer application, which was recommended so that fertilizer will not have impact on Striga emergence. In-vitro experiments have shown that nitrogen in form of ammonium or nitrate inhibits germination and extension of radicle length of Striga as a result of inhibition of production of chemical stimulants by host plants [51] . Studies have also shown a toxic effect of nitrogen on Striga development following attachment [52] . This research was targeting the cheapest Striga control method for resource poor farmers who cannot afford fertilizer for its control. It was therefore important not to make fertilizer to suppress the parasite in order to see the actual effect of Afayak as well as the Striga resistant maize varieties. Notwithstanding the insignificance across treatments, the maize varieties in integration with and without Afayak demonstrated potentials in yield in Striga endemic area, with sole Wang dataa recording the highest grain yield (1067 kg/ha) and Aburohemaa the lowest (153 kg/ha).
4. Conclusion
The results showed that, there was no relativity of Striga stress tolerance in terms of yield and yield components of the four maize varieties. All the entries efficiently tolerated the biotic stress of Striga and further supported growth and grain yield equally. There was reduced S. hermonthica seed bank production in the soil in both cropping systems. This has the positive implication of reducing the capacity of the S. hermonthica seed bank production in the soil. The performance of the maize varieties in integration with and without Afayak demonstrated a potential in yield in Striga endemic areas.
5. Recommendations
Considering the results of this trial, it is recommended that resource-poor farmers in Striga endemic areas can use the four tested Striga tolerant maize varieties (Omankwa, Wang dataa, Aburohemaa, and Bihilifa) mono crop or intercrop with Afayak to manage Striga hermonthica. The four Striga tolerant maize varieties (Omankwa, Wang dataa, Aburohemaa, and Bihilifa) can be planted sole or intercrop with Afayak to deplete Striga seed bank in Striga endemic areas in the Guinea Savannah zone of Ghana. However, it must be noted that to avoid replenishment of Striga seed bank in the soil, farmers should combine this method with cultural practices such as hand pulling or weeding of emerged Striga seedlings.