**1. Introduction**

*Brachiaria* (*Urochloa*) is a genus in Poaceae family commonly called brachiaria and grown for forage in Latin America, Asia, South Pacific, and Australia [1]. A widely grown species *Brachiaria brizantha* represents 85% of cultivated pastures in brazil alone [2]. In Africa, where they originate from, they are natural constituents of grasslands in eastern, central, and southern regions [3]. Brachiaria is recently identified as an ideal fodder that can improve livestock production in eastern Africa. This is due to its adaptability to low fertility areas, arid, and semiarid zones of sub-Saharan Africa [3]. There are several initiatives in the region aimed at promoting cultivation of brachiaria to support the emerging livestock industry [4].

A reduction of fall armyworm, (*Spodoptera frugiperda*) damage was recently observed in push-pull technology (PPT) plots as compared to farmers' practice [5]. Furthermore, desmodium enhances soil integrity. Further, *Brachiaria* cv. Mulato II is an important companion crop used as a trap plant in a push-pull technology (PPT, www.push-pull.net). Developed by International Centre of Insect Physiology and Ecology (*icipe*, Nairobi, Kenya), Rothamsted Research (Harpenden, UK), and national partners, PPT is a conservation agriculture system for integrated pest, weed, and soil fertility managemen<sup>t</sup> in crop–livestock farming systems [6,7]. The system involves intercropping the main crop, either maize *Zea mays L*. or sorghum *Sorghum bicolor* (L.) Moench with a fodder legume, silverleaf desmodium *Desmodium uncinatum* (Jacq.) DC., and surrounded with Napier grass, *Pennisetum purpureum* [8]. The climate smart variant uses drought tolerant green leaf desmodium, *Desmodium intortum* (Mill.) Urb., and brachiaria *B. brizantha* cv Mulato II as the border crop [9,10]. Desmodium releases chemicals that repel stemborers, while volatile chemicals from Napier grass or brachiaria attract the insects and their natural enemies. Through this chemistry, PPT significantly reduces the infestation of cereal stemborers *Busseola fusca* (Fuller) (Noctuidae) and*Chilo partellus* Swinhoe (Crambidae) [10–13]. Significant through nitrogen fixation improves soil organic content and conserves the soil moisture [14,15]. Root exudates of desmodium cause the abortive germination of a noxious weed striga *Striga haemonthica*, therefore providing additional benefits in weed managemen<sup>t</sup> [11]. On the other hand, brachiaria is a high-value forage crop that facilitates milk production and diversifies farmers' sources of income. A recent study shows that *B. brizantha* cv Piata has higher content of dry matter, crude protein, and organic matter than Napier grass [16]. Some species of brachiaria reduce emission of nitrous oxide from the soil through biological nitrification inhibition [17,18].

Agriculture is an economic mainstay in most developing countries of the tropics. It is characterized by smallholder mixed crop-livestock systems where rain-fed crops and livestock are raised on the same farm [19]. In sub-Saharan Africa (SSA), smallholder farming is a major source of food production and income, contributing up to 80 percent of food consumed [20]. Crops mostly cultivated in the region are, in order, maize, cassava, rice, sorghum, wheat, and millet, while livestock species include cattle, goats, and sheep [19]. However, production in the region is constrained by climate change related biotic and abiotic constraints, including pests and disease outbreaks, extreme weather conditions, among others. This poses a threat to food security and livelihood in communities dependent on agriculture [21,22]. More than 250,000 smallholder farmers in sub-Saharan Africa have used push-pull technology to manage stemborers, fall armyworm, noxious Striga weeds, and soil fertility, and to generate livestock fodder [23].

The value of brachiaria to African agriculture observed thus far can be optimized by addressing the current and foreseeable production constraints. Yet, the few genotypes that are commercialized in Africa were developed for the Americas and Australia thus a higher risk of pest and disease attacks, coupled with poor adaptability to local environments. There is, remarkably, a wide genetic variation in the genus *Brachiaria* [24] that can be exploited in breeding programs for locally adapted genotypes. Recent studies identified brachiaria genotypes that combine drought tolerance and moderate resistance to spider mites [25,26]. Among these genotypes, some are attractive to oviposition by stemborer moths while being detrimental to the larvae of the pest, thus valuable in push-pull technology [27]. These genotypes could be of value in the improvement of cereal livestock-based livestock productivity in sub-Saharan Africa in the current scenarios of increasing aridification and attacks by invasive pests, such as spider mite (*Oligonychus trichardti*). However, farmers' skills and knowledge can complement scientific research and their contribution through participatory approach is key in validating the potential of such genetic materials. Therefore, this study aimed at (i) assessing smallholder farmers' perception on benefits of brachiaria in cereal-livestock production, (ii) assessing brachiaria production constrains, and (iii) identifying farmer preferred brachiaria genotypes.

#### **2. Materials and Methods**

#### *2.1. Study Area*

Arid and semi-arid areas of western Kenya and the lake zone in Tanzania were selected because of their importance in cereal-livestock based farming systems. The study areas in Kenya included Homabay, Mbita, Bondo, and Siaya. The lake zone in Tanzania (hereafter referred to Tanzania LZ) included Tarime and Mwanza districts. Rainfall pattern is bi-modal, main season runs from March to August and the short season is from October to January. Farming systems in the regions are predominantly cereal/edible legume integrated with livestock [6]. These areas are historically hot spots for cereal stemborer, and most farmers have widely adopted push-pull technology (PPT) as a managemen<sup>t</sup> tool for the pest [5]. Further, the areas are characterized by extended periods of drought, which makes them conducive for the invasive spider mite, *O. trichardti.* Spider mite is the most important pest of brachiaria, a companion crop in PPT, especially during drier and hotter regimes [26]. These study areas are therefore ideal for assessment of farmers' experience and preference of brachiaria genotypes for use in cereal-livestock production.

## *2.2. Demonstration Plots*

One site per study area was selected for the establishment of the demonstration plots for six brachiaria genotypes. This was purposely done by ensuring that they occur in di fferent agro-ecologies, as follows; Homabay (Lower Midland 3), Mbita (Lower midland 5), Bondo (Lower midland 4), and Siaya (Lower midland 2) [28]. The lake zone sites in Tanzania included Tarime (high altitude plateau) and Mwanza (medium altitude plains) [29]. Brachiaria genotypes that were planted for evaluation were Piata, Xaraes, Marandu, ILRI 12991, ILRI 14807, and Mulato II (check). The candidate genotypes were selected from previous studies that tested drought tolerance, adaptability to a range of environments, and resistance to spider mite in brachiaria [25,26]. Furthermore, Xaraes, Piata, and Marandu are suitable for egg laying by the lepidopterous stemborer *Chilo partellus* and are, therefore, suitable companion plants in PPT [27]. Each plot measured 5 × 5 m with plant to plant and row to row spacing of 0.5 m. Diammonium phosphate (DAP) was applied as basal fertilizer at a rate of 60 kg/ha, and nitrogen in the form of calcium ammonium nitrate (CAN), at a rate of 60 kg/ha as at top dresser four weeks after planting. The plots were kept weed free by hoe and hand weeding and pesticides were not applied to allow natural infestation and the development of spider mites.

#### *2.3. Sampling Procedures*

The selection of respondents to participate in study at specific trial sites followed a multistage sampling procedure. Firstly, farmers who practiced climate smart PPT and were within the study areas were selected for the study. This was done by generating a checklist of all farmers who practice climate smart PPT with help of village elders and frontline extension sta ff. Thereafter, a semi-structured questionnaire was used to identify willing respondents for participatory evaluation of di fferent brachiaria genotypes grown in the demonstration plots.

#### *2.4. Data Collection*

The study used semi-structured questionnaires that were administered through individual interviews and focus group discussions (FGD). The questionnaires were pre-tested before implementation. Individual response questionnaire assessed farmers' socio-economic characteristics (e.g., age, gender, and education), farm characteristics (farm size, tenure system, size of land under brachiaria, and uses of brachiaria). Farmers' perceptions on whether brachiaria was beneficially in controlling cereal pests, for sale, as livestock feed, for soil conservation, etc., was sought. Production challenges, including access to planting materials, planting, crop management, harvesting, and hay making was recorded. The questionnaire also assessed farmer experience with pests and diseases of brachiaria. This was captured by asking the respondents whether they had noticed the infestation of the red spider mites and sorghum shoot flies; severity of infestation; and, what they did to cope with the pests. Rating of the seriousness of the pest was based on a four-point Likert scale, where 0 = no problem, 1 = moderate problem, 2 = severe problem, and 3 = very severe problem. Farmers were asked whether they are aware of any other brachiaria genotypes and whether they had planted them on their farms. Thereafter, farmers evaluated the di fferent brachiaria genotypes in demonstration based on hairlines, leaf size, leaf softness, number of shoot tillers, plant spread, plant height, seed setting, resistance to spider mites, and biomass yield where their responses were based on a scale of 1 (poor), 2 (fair), 3 (good), and 4 (excellent). However, for each trait, the number of the highest score i.e., 4 (excellent) was used to

compare the genotypes. Evaluation based on sorghum shoot fly damage was not done, since there was no infestation; this is because the genotypes were raised from root splits. Sorghum shoot flies are known to attack young seedlings, especially when grown from the seed. During the assessment, the genotypes were given numbers instead of their actual names to reduce bias in ranking of popular genotypes. Farmers were finally asked to select the best brachiaria genotype. To back up individual interviews, focus group discussions were conducted. Farmers were encouraged to use a language that they were most familiar with and discussions were led by a member of the research group who spoke their language. Similarly, the discussions covered the benefits of brachiaria in climate smart PPT, production constraints, including important pests (spider mites and sorghum shoot flies), and farmer preference of different brachiaria genotypes. Other aspects that were covered in the FGD included willingness of the farmers to try other brachiaria genotypes on their farms and the criteria used for selecting a candidate genotype.

## *2.5. Data Analysis*

Descriptive and comparative statistics (means, percentages, and cross tabulations) were used in data analysis. Analysis of variance, F-test, and Pearson's product moment correlation coefficient (chi-square test) were used to test for significance of differences in various responses and study areas. Computation was done using statistical package for SPSS version 17 software (IBM Corporation, Armonk, NY, USA).
