**3. Results**

#### *3.1. Farmer Socio-demographics and Farm Characteristics*

A total of 223 respondents participated in individual interviews. In general, 49% of the respondents are male. There was no significance (*p* = 0.05) in variation between districts (Table 1). Their age categories varied significantly (*p* = 0.01) between the study areas. The majority were between 41–50 years (31%), followed by 51–60 years (22%), the elderly >61 (19%), 31–40 (15%), and the least being the youth between 20–30 years (8%). Age category 41–50 years formed the highest percentage in all study areas, except in Siaya, where the majority were between 51–60. The lowest in population in all areas were the youth (20–30 years), implying that farming in the region is predominantly practiced by the older and the elderly farmers. Education levels of the respondents varied significantly across the study areas. On average, 50% of the farmers had attained primary education, 31% had secondary education, a few (9%) had post-secondary education, while those with none-formal and no education at all were the least, each comprising of 4% of the respondents. Literacy at post-secondary level was the highest in Bondo (15%), while illiteracy (no education) was highest in Siaya (13%) (Table 1).

Farmers rented an average of 1.5 acres land for farming. This varied significantly across the study areas ranging from one acre (Homabay) to two acres (Mbita). The average farm size owned was 3.5 acres, and it varied significantly (*p* = 0.01) from two acres (Siaya) to 5.2 (Homabay). The average land size under brachiaria as components of push-pull was 0.17 acres, this however did not vary significantly across the study areas (Table 1). Besides the brachiaria that forms a component of push-pull, 22% of farmers planted the grass as pure stands; 32% in Homabay, 3% in Mbita, 19% in Bondo, 39% in Siaya, and 11% in Tanzania. The majority of the farmers (93%) kept livestock that included cattle (improved and local), goats (improved and local), and sheep. The type of livestock mostly kept was local cattle with a mean of 2.9, followed by local goats (1.94), sheep (1.65), improved dairy cattle (0.72), and improved dairy goats (0.56). Variation across study areas was not significant for all animals, except sheep (*p* = 0.01). The highest number of sheep was recorded in Homabay (2.92), while the lowest was in Siaya (0.64).



#### *3.2. Benefits of Brachiaria*

The uses of brachiaria varied significantly across the study areas, except for those who exchanged the grass for milk (mean = 5%). Being a component of a pest managemen<sup>t</sup> strategy, approximately 38% (mean) of farmers agreed that brachiaria reduced damage caused by cereal pests (Figure 1). Approximately 36% of the farmers considered the grass as a valuable fodder for livestock. About 29% believed that it controls soil erosion, while a few (17%) also sold the grass, others exchanged it for milk (Figure 1).

**Figure 1.** Main uses of brachiaria by the small-holder push pull farmers.

#### *3.3. Constraints to Production of Brachiaria*

Farmers gave their opinions on challenges faced in accessing planting materials, during planting and managemen<sup>t</sup> (Table 2). The opinions varied across the study sites. The major production constraint perceived by both men and women was attacks by arthropod pests (49.2% and 63.1, respectively). This was followed by disease infestation (44.4% of men and 30.6% of women). The difficulty in handling (prickly hairs) was rated as the third constraint by women (24.7%) while unavailability of seed in agrovets was rated third by men (16%). However, farmers were specific about the most important arthropod pests. Spider smites were reported to have been observed on own farms by 50.8% of men and 63.1% of women, while sorghum shoot flies had been observed by 58.1% of men and 67.9% of women. The seriousness of spider mites was mostly perceived by both men and women as moderate problem (63.4% and 44.7%, respectively) to severe problem (21% and 33.9%, respectively). Sorghum shoot flies were also regarded by both men and women as a moderate problem (58.8% and 52.2%, respectively) to severe problem (15.8% and 22.4%, respectively). Further, more farmers (55.6% men and 63.2% women) had observed spider mites on other farms. On the other hand, more men (50.3%) had not observed sorghum shoot flies on other farms, while more women (53.1%) have observed.

**Table 2.** Perception on brachiaria production challenges disaggregated by gender.


#### *3.4. Farmer Evaluation and Selection of Brachiaria Genotypes*

Farmers assessed the six brachiaria genotypes based on the following criteria; leaf hairlines, leaf size, leaf softness, number of shoot tillers, plant spread, plant height, seed setting, resistance to the spider mites, and visual estimation of biomass yield. Mulato II ranked the highest in hairlines, leaf softness, and tillers, while Xaraes had the highest numbers for plant height, resistance to spider mites, and biomass yield (Figure 2). Sorghum shoot flies attack the crop at the seedling stage, especially when seed is used as a propagation material. Due to unavailability seeds, we used root splits in our study, further, the crop was evaluated at maturity. Therefore, farmers could not evaluate the materials based on resistance to Sorghum shoot flies. Figure 3 presents the farmer selection of different brachiaria genotypes. Generally, the majority (41.2%) of the farmers preferred Xaraes, followed by Mulato II (25.6%) and Piata (20.4%).

**Figure 2.** Average number of farmers recorded for the highest score 4 (excellence) for different traits of brachiaria.

**Figure 3.** Means for number of farmers (%) for each brachiaria genotype across all the districts. Bars represent standard error of the means.
