Response of improved Brachiaria (Urochloa) grass cultivars to foliar diseases and their agronomic performances in Rwanda

Diseases have emerged as one of the major challenges of Brachiaria production in Africa. Nine Brachiaria cultivars were evaluated for leaf rust, leaf spot and leaf blight diseases and agronomic performances in two agro-ecological zones of Rwanda. The relationships between agronomic traits and area under diseases progress curve (AUDPC) were determined using Pearson correlation analysis. Cultivars differed significantly (p ≤ 0.05) for their response to these three diseases. Basilisk, Marandu, MG4, and Xaraes exhibited moderately resistant to resistant response to all three diseases, but Cayman and Cobra were susceptible to leaf rust. Site × cultivar × harvest interaction was significant for diseases and agronomic parameters (p ≤ 0.05). Cultivars also differed significantly for biomass production and dry matter content (p ≤ 0.05). The highest biomass producers were Marandu and Xaraes, and Cayman, Cobra and Piata had highest dry matter content. The AUDPC for leaf rust and leaf spot had negative and significant correlation with biomass yield. Our study concludes a satisfactory level of resistance in Basilisk, Marandu, MG4 and Xaraes to all three foliar disease in Rwanda. We recommend routine surveys for emerging and re-emerging Brachiaria diseases and studies to develop effective management measures against Brachiaria diseases.

Brachiaria (Urochloa) grass is one of the most important tropical forages distributed in the tropical and subtropical regions of eastern and western hemispheres (Renvoize et al. 1996). Brachiaria grass possess several characteristics of agricultural and environmental significance, such as high biomass yield, nutritious to livestock, drought and shade tolerance, soil fertility improvement, high nutrient use efficiency, mitigation of the climate change adversity and effective bioagents for pests and parasitic weed management (Subbarao et al. 2009;Maass et al. 2015;Khan et al. 2016). The genus Brachiaria consists of approximately 100 species, of which seven perennial species of African origin have been exploited for fodder production, and they have been cultivated in Asia, Australia, the South Pacific, and South America at various scales (Stür et al. 1996;Jank et al. 2014). Brachiaria species are a common and valuable constituent of natural vegetation in East Africa (Boonman 1993), but their use as sown pasture for animal production is very recent in the region (Maass et al. 2015;Njarui et al. 2016). A broad adaptation, excellent animal performances and high biomass yields are among the major factors that promote a wider and rapid adoption of Brachiaria grass across the tropics and subtropics. However, the productivity of Brachiaria grass is affected by different abiotic and biotic factors, including diseases causing high yield losses (Nzioki et al. 2016;Hernandez et al. 2017). Diseases of Brachiaria grass, their symptoms, geographical distribution, and management options have been reviewed by Valerio et al. (1996). These diseases have negative impact on livestock productivity, because they reduce forage yields and quality.
Brachiaria grass is one of the preferred forages by livestock keepers in Rwanda and in other East African countries (Mutimura and Everson 2012;Maass et al. 2015). As mentioned earlier, diseases are among the major biotic constraints of Brachiaria grass production, and diseases, such as leaf rust, leaf spot, and leaf blight are reported to cause economic loss (Lenné and Trutman 1994;Miles et al. 1996;Rao et al. 1996;Alvarez et al. 2014;Maass et al. 2015). For instance, leaf rust can cause up to 100% yield loss, reduces leaf crude protein content to between 49 and 53% and subsides availability of other nutrients (Lenné and Trutman 1994). Similarly, leaf blight reduces forage biomass yield by approximately 50% in the tropics (Alvarez et al. 2013). Recent studies have shown widespread distribution of leaf rust, leaf spot, and leaf blight diseases in Kenya and Rwanda (Nzioki et al. 2016;Uzayisenga et al. 2020). Therefore, sustainability of Brachiaria grass production in Africa relies on how well these diseases are managed.
Many pastures including Brachiaria grass are considered low-value crops, often cultivated in large acreage and management of diseases using chemical is too costly and

Introduction
Open Access article distributed in terms of the Creative Commons Attribution License [CC BY 4.0] (https://creativecommons.org/licenses/by/4.0) not safe for livestock and environment. Therefore, disease management efforts in Brachiaria grass should focus on low-cost control measures like host-plant resistance that is effective, economical, easy to apply and safe for environment. Currently, some improved Brachiaria grass cultivars are available to address major production challenges like biomass yield, nutritive quality, drought tolerance, and pests and disease management (Lenné and Trutman 1994;Alvarez et al. 2014;Maass et al. 2015). For example, cultivars Mulato and Mulato II were developed for spittlebug resistance, high forage yield and nutritive quality (Miles et al. 2004;Argel et al. 2007). Cultivar Cayman was developed for water logging tolerance (Pizarro et al. 2013), and some Brachiaria hybrids were developed for foliar blight resistance (Alvarez et al. 2014).
Demand for improved Brachiaria grass is high in Sub Saharan Africa. Therefore, many livestock development initiatives implemented by National Agricultural Research System (NARS) institutions in Sub-Saharan Africa, international organizations and development agencies have been promoting Brachiaria grass in the continent as a nutritious and climate resilient forage. These programmes currently rely on a few improved cultivars initially developed for South America with extremely narrow genetic base (Keller-Grein et al. 1996) Mutimura and Ghimire 2021). In 2013, eight improved Brachiaria cultivars were introduced and evaluated for adaptation and biomass yield in different agro-ecological zones . These cultivars were successfully integrated into mixed crop-livestock system that subsequently improved forage availability and livestock productivity (Mutimura et al. 2016;Mutimura et al. 2018).
All improved Brachiaria grass cultivars that are introduced and promoted in Africa were developed in South America and Australia. Some of these improved cultivars have shown broader adaptation, excellent agronomic performance and high livestock productivity stimulating high demand for Brachiara grass by farmers in Africa. However, many of these cultivars are susceptible to diseases, such as leaf rust, leaf spot, foliar blight, and ergot in Kenya and Rwanda (Nzioki et al. 2016;Uzayisenga et al. 2020). Therefore, expansion of Brachiaria acreage without proper disease management measures will increase chances of disease outbreaks and crop failure. The cultivation and scaling up of these improved cultivars over large geographical ranges in Africa requires some consideration of the existing and emerging diseases (Maass 2015), accordingly warranting routine evaluation of the existing and new cultivars against diseases. This study evaluated the response of nine improved Brachiaria grass cultivars to three major foliar diseases and assessed the effects of the foliar diseases on the agronomic performance of those cultivars in Rwanda.

Field experimentation
Nine improved Brachiaria grass cultivars were evaluated in four replicates using a randomised complete block design (RCBD) in two sites. In addition to the main treatments, a leaf blight susceptible cultivar, Mulato (Argel et al. 2007;Alvarez et al. 2014), was planted as disease spreader rows four weeks prior to planting of test cultivars to check whether the disease was present naturally in the test sites and to trap early inoculum for the disease development. The test cultivars were surrounded by two spreader rows of Mulato planted at 50 cm between rows and 25 cm between plants. Each test cultivar was planted on 3.5 m row accommodating 14 plants per replicate at a spacing of 25 cm within rows and 1 m between rows, and a 2 m spacing was kept between replicates ( Figure  1). Planting was done using minimum of two rooted tillers per hill. A basal dressing of cattle manure (10 t ha −1 ) and NPK17-17-17 (100 kg ha −1 ) were applied at the time of planting in the top soil (0-30 cm depth) in each planting hole, and urea was top-dressed two weeks after planting in rows (broadcast at 5 cm radius at each tiller) at the rate of 50 kg N ha −1 . Irrigation and weeding were performed manually as required. All test cultivars in the experimental plots were subjected to standardisation cut at 5 cm above the ground level four weeks after planting to stimulate tillering and uniform regrowth. For each cultivar, six stools showing uniformity in appearance and growth were selected in each replication and tagged for assessing diseases and agronomic parameters. Border effect was eliminated during data collection by excluding at least one stool from either side of a row. The field experiments covered three consecutives growing seasons: March to July 2019 (wet to semi-dry season characterised by high rain intensity, but shorter rains), August to December 2019 (dry to wet season, characterised by long rain-patterns/ distribution with medium rain intensity), and January to May 2020 (wet season characterised by shorter rains with high intensity).

Assessment of disease incidence and severity
Disease incidence and severity were recorded every four weeks after the standardisation cut up to the 20th week for the first, second, and third seasons. Five assessments were made for each of the three consecutive seasons. Disease incidence was assessed on six tagged stools per cultivar in each replication and was determined as the number of stools showing disease symptoms, then converted to percentage of the total number of assessed stools, i.e. samples collected from each replicate. On the same stools, severity of leaf rust, leaf spot, and leaf blight was assessed using the disease rating scales described in   area under disease progress curve was calculated from severity data collected over the five different time points in each season, as described by Shaner and Finney (1977).
The response of the test cultivars to leaf rust was determined based on infection types and the area under the disease progress curve (AUDPC). The leaf rust infection types were recorded at eight-week-old stools for all three seasons (two harvests in 2019 and one harvest in 2020) using a five-category scale, namely immune, resistant, moderately resistant, moderately susceptible, and susceptible (Roelfs et al. 1992). Similarly, the response to leaf spot and leaf blight was determined based on AUDPC data where lower AUDPC values correspond to resistance and higher AUDPC values correspond to susceptibility. The AUDPC values <3 500 correspond to resistant, values between 3 500 and 4 500 correspond to moderately susceptible, and the values >4 500 correspond to susceptible reaction (Magar et al. 2015;Pantha et al. 2017;Kumari et al. 2018).

Evaluation of agronomic parameters
The data on plant height and number of tillers per stool were taken every four weeks following the standardisation cut for each harvest, i.e. growth period/season. Plant height was recorded from the base of the stool up to the longest leaf of every tagged stool (Rayburn and Lozier 2007). The dry matter (DM) yield and dry matter content were recorded for each harvest. Harvesting was carried at 20 weeks' interval all through the seasons. For determination of dry matter content, 200 g subsample from fresh biomass was put in paper bags, and oven dried at 105 °C for 24 hours. The total dry matter yield (kg ha −1 DM) and the percentage dry matter were calculated as described by Wassie et al. (2018) and Oliveira et al. (2019).

Meteorological data
The rainfall and temperature data during the experimental period from March 2019 to May 2020 for Gashora and Rubona experimental sites was obtained from the Rwanda Meteorology Agency (RMA 2020). Thirty-year average data available at https://www.besttimetovisit.com.pk/ were used for any missing monthly rainfall and temperature data.

Statistical analyses
Data on disease parameters (incidence, severity and AUDPC) and agronomic traits (plant height, number of tillers, dry matter yield and the dry matter content) was subjected to analysis of variance (ANOVA) using GenStat for Windows 20th Edition (VSN International 2019). To account for the overall trend of disease incidence, disease severity, plant height, and number of tillers, data were subjected to general linear model predictions using repeated measures (Littell et al. 1991). All data were analysed at the cultivar × site × harvest/season interaction level and then presented based on the significance of their interactions (p ≤ 0.05). The means of disease and agronomic parameters were compared by Least Significant Difference (LSD) mean separation test at 0.05 probability level. Because the data were analysed at cultivar × site × harvest/season interaction, all cultivars, sites, and harvests/seasons were considered together to calculate Pearson correlations that allowed to illustrate the relationships among the agronomic traits (plant height, number of tillers, dry matter yield) and the AUDPC for leaf rust, leaf spot and leaf blight diseases.

Meteorological data at experimental sites
Monthly rainfall and monthly average temperature for Gashora and Rubona are presented in Figure 2. Sites differed distinctly in terms of annual rainfall and monthly rainfall differed between seasons ( Figure 2). In all three seasons, Gashora had higher mean monthly temperature than Rubona (first season = 22.7 °C vs 20.1 °C, second season = 23.4 °C vs 20.5 °C, and third season = 21.0 °C vs 20.2 °C). However, the differences in monthly temperatures between the sites were minimal for all three seasons.

Responses of improved Brachiaria cultivars to foliar diseases
The effect of cultivar, site, harvest, and interaction of cultivar by site and harvest was significant for leaf rust incidence, severity and AUDPC (p < 0.001; Table 3). The leaf rust incidence among the cultivars for all sites and harvests together ranged from 31.9% to 100.0% with the lowest incidence in Basilisk at Gashora in the third harvest and the highest incidence in Cayman, Mulato II and Cobra, mostly at Rubona in all harvests. Similarly, leaf rust severity ranged from 5.4% to 81.8% with the lowest severity in Basilisk in the third harvest at Gashora and the highest in Cayman in the third harvest at Rubona. The AUDPC for leaf rust was the lowest in Basilisk (754) and MG4 (949) in the third harvest at Gashora and the highest in Cayman (10 177)  There was significant effect of all treatments (except for site on incidence and AUDPC) and their interactions on incidence, severity and AUDPC of leaf spot disease (p < 0.001; Table 4). The incidence of leaf spot was the lowest in Cayman at Rubona in the first harvest and the highest in Humidicola at Rubona in the third harvest. The severity of leaf spot was the lowest in Cayman at Rubona in the first harvest and the highest in Humidicola at Rubona in the first harvest. The AUDPC for leaf spot ranged from 25 to 5 996, where Cayman had the lowest values at Rubona in the first harvest and Humidicola had the highest value at Gashora in the first harvest. All cultivars across the sites and harvests showed resistant reaction to leaf spot, except Cayman, Humidicola, and Piata, which were moderately susceptible/susceptible.
All treatments and their interactions had significant effect on the incidence, severity and AUDPC of leaf blight (p ≤ 0.001; Table 5). Cultivar Humidicola had the lowest incidence and severity of leaf blight in the first harvest at Gashora. Cultivar Cayman had the highest leaf blight incidence and severity in the third harvest at Rubona. Humidicola in the first harvest at Gashora and Cayman in second harvest at Rubona had the lowest values of AUDPC (0.0%) for leaf blight disease, but Cayman had the highest value (2 317) in the third harvest at Rubona.

Evolution of disease incidence at Gashora and Rubona sites
Results of the evolution of foliar diseases from four weeks to the 20th week showed that incidence of leaf rust, leaf spot and leaf blight increased with the number of weeks after standardisation cut and/or harvest (Figures 3, 4 and 5). Higher incidence of leaf rust disease ranging from 80-100% was observed for moderately susceptible and susceptible cultivars at Rubona at four weeks for all harvests considered together. It reached 100% at eight weeks and 12 weeks for moderately resistant cultivars. Over time, leaf rust incidence was the lowest in moderately resistant cultivars at Gashora (Figure 3). The leaf spot disease showed high diversity of cultivar responses throughout the crop growth periods (Figure 4). Leaf blight incidence was low in all cultivars until 16 weeks then spiked at 20 weeks. Exception at 16 weeks, leaf blight incidence was consistently higher at Rubona than Gashora ( Figure 5).     and 20.1 t ha −1 (Marandu in the first harvest at Rubona) and 20.2 t ha −1 (Xaraes in the third harvest at Gashora). Cultivar Piata had the highest dry matter content in third harvest at Rubona. Cultivar Humidicola had the lowest dry matter content in the third harvest at Gashora.

Correlation between disease intensity and agronomic parameters
Pearson's correlation analysis revealed significant differences (p ≤ 0.05; Table 7) between agronomic parameters (plant height, tiller number, dry matter yield and dry matter content) and the AUDPC for leaf rust, leaf spot and leaf blight diseases. There was a positive significant correlation of plant height and number of tillers with dry matter yield. Plant height was negatively correlated with number of tillers with no significance. A negative and significant correlation was found between plant height and AUDPCs for leaf rust and leaf blight diseases. A negative and significant correlation was observed between AUDPCs for leaf rust and leaf spot diseases and dry matter yield (Table 7).

Discussion
Diseases are one of the major biotic constraints limiting fodder production and qualities of Brachiaria grass with a negative impact on the forage availability and livestock production. Many diseases caused by fungi, bacteria and viruses have been reported on Brachiaria grass (Lenné 1990;Valerio et al. 1996;Nzioki et al. 2016;Uzayisenga et al. 2020). Some of these diseases have the potential to cause complete failure of the crop. Therefore, the sustainability of Brachiaria grass production in the tropics and subtropics depends on how effectively and economically these diseases are managed. Amongst various disease management methods, the use of resistant cultivars is an effective and least cost control measure  Values with the same superscript letters within the columns are not statistically different at p ≤ 0.05 ** Correlation is significant at the 0.01 level (two-tailed), * Correlation is significant at the 0.05 level (two-tailed), ns = not significant at the 0.05 level (two-tailed), AUDPC = Area under disease progress curve. All cultivars, sites and harvests were considered together that can be easily adopted by farmers, including resourcelimited smallholder livestock keepers from the developing countries. In this study, we evaluated nine improved Brachiaria grass cultivars against three major foliar diseases, namely leaf rust, leaf spot and leaf blight, in two sites, each representing different agro-ecological zones of Rwanda for three consecutive harvests. In addition, each cultivar's agronomic performances were documented and relationships between diseases and selected agronomic traits were determined. The Brachiaria grass cultivars evaluated in this study reacted differently to leaf rust, leaf spot and leaf blight diseases. Moreover, the interaction effect of cultivar × site × harvest was obvious on the foliar disease development. These results could be combined effects of (i) difference in the genetic background of cultivars, (ii) variation in the virulence level of pathogen, and (iii) the biophysical characteristics of experimental sites that support or limit the disease development (Agrios 2005). The observation that genotypes (e.g. Basilisk and Cobra) expressed symptoms of all three diseases may indicate the absence of host immunity against a specific fungus associated with each disease.
The effect of Brachiaria cultivar, experimental site and harvest, and interaction of cultivar × site × harvest on the foliar disease development were evident. Earlier studies reported the presence of susceptible and resistant traits in Brachiaria grass germplasms to leaf rust and foliar blight (Torres and Trutmann 1991;Alvarez et al. 2014), implying varying level of susceptibility to diseases among Brachiaria cultivars (Kamidi et al. 2016).
The evaluation of B. brizantha, B. decumbens, B. dictyoneura and B. humidicola accessions against leaf rust in two regions of Colombia indicated susceptibility of all cultivars, except B. decumbens, in both regions (Torres and Trutmann 1991). Surveillance of Brachiaria grass diseases in Kenya and Rwanda revealed differences in the level of disease development in a cultivar between study sites and season (Nzioki et al. 2016;Uzayisenga et al. 2020), ratifying the influence of the biophysical characteristics of test locations on the disease development.
Environmental factors played important roles in the initiation and development of diseases. Though the differences in mean temperature between experimental sites were minimal, there were noticeable differences between sites for rainfall. A significant variation in the incidence and severity of Brachiaria foliar diseases in different districts of Rwanda with diverse agroclimatic conditions has been reported previously (Uzayisenga et al. 2020). A high level of leaf rust development at Rubona in this study could be associated with moist conditions, because of high and well distributed rainfall throughout the crop growing periods. Rubona received 1.65 times greater rainfall than Gashora during the wet to semi-dry season of 2019, as well as in the wet season of 2020. The effect of climatic variables, such as temperature and humidity on wheat rust disease development had been reported by Barrera et al. (2012) and Sandhu et al. (2017). The high incidence and severity of leaf rust and blight at Rubona could have been favoured by high rainfall regime that is associated with moist environment. On the contrary, the relatively dry weather at Gashora might have supported leaf spot disease development. High incidence and severity were registered as the week advanced and to some extent in the subsequent harvests. These findings suggest that continuous presence of a perennial crop like Brachiaria grass in the field favoured pathogen population build-up over seasons and occurrence of the disease early in the season affecting crop performance and yields. In most cultivars, leaf rust incidence declined after 16 weeks, which might be because of the removal of rust spores, and climatic conditions, including rain. Another probable reason is adult plant resistance to rust disease.
The Brachiaria cultivars Basilisk, Marandu, MG4, and Xaraes were moderately resistant to all three foliar diseases and they had higher biomass yields than other cultivars confirming their suitability for cultivation in wider geographical regions. The low level of leaf-blight-disease development in all cultivars could be because of a low natural disease pressure in experimental sites and/or presence of intrinsic resistance in some cultivars (Kelemu et al. 1995;Alvarez et al. 2014) that merits further investigation. Alternative disease management options for rust-susceptible cultivars that produced high dry matter content (e.g. Basilisk and Cobra in this study) can be another area for further investigation, because this trait has a significant role in feed availability and improving livestock productivity.
Significant negative correlation between AUDPC of foliage diseases and dry matter yield indicate negative impact of these diseases on herbage yield of improved Brachiaria cultivars. This might be attributed to the reduction of the photosynthetic area in the diseased leaf tissues as reported for wheat diseases (Kandel et al. 2014;Lamsal et al. 2017;Pandey et al. 2018), as well as pathogen's dependence on host for nutrients and water. We observed significant and negative correlation between leaf rust and foliar blight and plant height conferring the negative effect of diseases on plant growth and development. Other studies have also reported negative correlation between wheat spot blotch disease and plant height (Joshi et al. 2002;Rosyara et al. 2009;Neupane et al. 2013).
The use of Brachiaria grass for pasture improvement and ruminant feeding in Africa started through the introduction of improved cultivars and hybrids mostly from the South America (Maass et al. 2015;Njarui et al. 2016). Within a very short period, Brachiaria grass has become a forage of choice for many livestock farmers on the continent to address challenges, including shortage of livestock feeds, poor nutritive value of local forages, extended and frequent drought, and declining productivity of extensively cultivated Napier grass Negawo et al. 2017). The importance of Brachiaria grass as an additional forage option in Sub Saharan Africa to improve livestock productivity through improving supply of quality herbage have been established recently (Mutimura et al. 2016;Njarui et al. 2016;Mutimura et al. 2018).
There are reports on attack of improved Brachiaria cultivars by multiple diseases caused by fungi, bacteria and viruses affecting the yields and qualities of forage and seed crops (Nzioki et al. 2016;Kamidi et al. 2016;Uzayisenga et al. 2020). There is a need to develop new cultivars with improved disease resistance qualities, exploring genetic diversity available in genebanks and natural populations preferably from East Africa, which is considered a centre of diversity of the genus Brachiaria (Keller-Grein et al. 1996).

Conclusion
This study revealed a wide variation in nine improved Brachiaria grass cultivars to leaf rust, leaf blight and leaf spot diseases in Rwanda. Besides variable responses to diseases, cultivars were also different for important agronomic traits, such as plant height, tiller numbers, biomass yield and dry biomass content. This study identified Basilisk, Marandu, MG4, and Xaraes as moderately resistant/resistant to all three foliar diseases and had high biomass yields. These cultivars are more appropriate for upscaling Brachiaria grass in wider geographical regions in Rwanda and neighbouring countries. There are reports available on the short life of disease-resistant cultivars of many agriculturally important crops; therefore, activities, such as periodic disease surveys, breeding new disease-resistant cultivars and development of other disease management options, should be given high priority for effective and sustainable management of Brachiaria grass diseases.