Abstract

This paper analyses perceived constraints encountered by farmers while sourcing appropriate information through various communication channels relevant to the adoption of weed management technologies. It further examines the values farmers attach to some pasture weeds that have contributed to the variable uptake of weed management technologies. The study was stimulated by the apparently low adoption of pasture weed management technologies in predominantly livestock dependant communities. In many cases, governments and industry employ a variety of communication media when distributing information relating to improved agricultural practice.

Single subject interviews were carried out with 165 participants from two ethnic communities within the south-eastern Kenyan rangelands to explore the uptake of knowledge relating to weed management technologies. An assessment was made of the techniques frequently employed and the obstacles presented to the uptake of new technology. In addition, an assessment of the benefits of some pasture weeds in farmers’ perspectives was assessed. Results indicate that extension agents were largely ineffective in the dissemination of weed management technologies in both communities as was the use of mass media. Other traditional information dissemination mechanisms such as using community decision leaders, neighbours and seminars were less effective than anticipated. Regardless of communication channels and sources of information, it was also found that pasture weeds were valued by farmers in both divisions as livestock feeds and woodfuel. Farmers in Mashuru placed importance on some pasture weeds as a source of construction material and medicinal plants. These mixed reasons for non-uptake of new technologies are discussed and recommendations made for future programs, including the need to improve the skills and resourcing of extension services across cultural settings and production system.

Media summary

A research project concluded that the adoption of pasture weed management technologies in Kenyan rangelands would be enhanced through consistent interaction between farmers and extension agents, effective communication channels and recognition of values farmers attached to some weeds.

Key words

Inter-personal, communication, production system, participatory, dissemination

Introduction

The transfer of technical information from the source to the desired audience requires appropriate channels of communication. This process of information transmission is utilised in agricultural extension to enhance the adoption of new technologies. According to Marra (2001) the adoption of technology involves the gathering of information in order to make a decision about its overall attractiveness and the level of risk involved in taking it up. This process involves attitudinal and behavioural changes of individuals towards new practices, which are influenced by the type of communication channels used. However, not all channels of communications are preferred equally by farmers. Preference depends largely on how farmers have been sensitized to a particular channel (Ekoja 2002).

In order for a message to be received, it needs to move from the message source through appropriate dispersal channel. While Rogers (1995) defined the message source as an individual or institution that originates the message, Marra (2001) explained that the message source can be extension agents, successful neighbours, other farmers, opinion leaders and organized groups entrusted with transferring technologies to potential adopters at different levels and circumstances. Dispersal channels are methods for transmission of information from the source to the receiver. Most commonly used channels of communication include mass media (radio and television), print media (pamphlets, brochures, newspapers, labels and magazines) and inter-personal media (seminars, demonstrations, field days, exchange visits, agricultural shows) (Rogers 1995).

Several studies have explained the role of extension in agriculture. It has been noted that while the extension agencies provided advice to farmers on how to overcome technological barriers, the decision to adopt or reject a technology was vested in the farmer (Ban et al. 1988). In addition, there are variations among countries or regions of the world in the extension approach used in the dissemination of agricultural technologies. In Kenya, the extension strategy used include “barazas” (group meetings), meetings with individual ranchers, field visits to farms and range research stations, training workshops in training centres, dissemination of information through radio programs and television, and the use of posters and pamphlets (Maranga 1998). It is informative to note, however, that extension in Kenya has been supported strongly in policy documents, yet physical and logistic support has been weak (Barrow 1991). As a result, the awareness and promotion of agricultural technologies have been hampered.

Extension agents are an important link between resource agencies and the community and are particularly important for agricultural technology adoption. Extension agents often provide a better “resource” for farmers, particularly decision makers, to share accumulated experience and views regarding new approaches and agricultural practices than other media (Chamala 1987b). This may be attributed to the influence of these influential group members in moulding beliefs, attitudes and perceptions towards adoption of new practices and technologies (Muli 1997). They are also key players in designing communication strategies and messages for capacity building programs (Ban et al. 1988). However, Crouch (1981) noted that extension agents spent more than half of their time in administrative activities not related to the client community, thus diluting the impact of their services.

Not all authorities agree that the “trickle down” theory is applicable to current farming systems, and some would argue that the “top-down” approach is perhaps more appropriate for resource rich communities (Rogers 1995 and Chamala 1987b). This may vary with different social systems as associations occur with ‘like’ ethnic, religious, social or educational groups (Chamala 1987b). However, innovations supported by opinion leaders in a social system are more likely to be adopted by followers irrespective of their risks and uncertainties. This is because, opinion leaders in most cases are respected people in a social system, more exposed to sources of information and have higher status than other members (Rogers 1995). For example in a Brazilian study of agricultural adoption, Rogers (1995) characterised opinion leaders as people with larger farms and large livestock herds who adopted new practices more often than their followers. To complement the extension agents and opinion leaders, successful neighbours can also become role models for others in the community in relation to trialled technologies (Foster 1995).

While the ‘face to face’ value of extension services has made it a popular choice in recent years, mass media has increased the reach and credibility of the messages of management agencies. Rogers (1995) found magazines, container labels and bulletins more effective in influencing choices for chemical weed control than personal contact. Therefore, failure to effectively use a range of media may dramatically reduce the effectiveness of campaigns. Inadequate use of mass media in Thailand was believed to be responsible for less than 4% of farmers in the rural areas accessing new agricultural technical information (Pontius 1983). Furthermore, negative and non-relevant messages and language presented in the media have contributed significantly to non-adoption of technologies in developing countries (Lingamneni 1981).

Therefore the objective of this study was to investigate, by means of interviews with farmers, the effectiveness of the various communication media used in enhancing the adoption of weed management technologies in the south-eastern Kenyan rangelands. Many studies have identified extension agencies, mass media and neighbours (decision formers/opinion leaders) as important components in the facilitation of increased adoption rates of new agricultural practices (Guerin et al. 1994). No study has been undertaken to date to determine the influence of sources of information and channels of communication in the adoption of weed management technologies in the south-eastern Kenya rangelands. This study attempts to expand on previous research in extension in order to ascertain the effective means of extension in the south-eastern Kenyan rangelands in promoting pasture improvement and consequent increase of livestock productivity. Further, it attempts to distinguish among the extension strategies employed in two different cultural regions and production systems.

Research site

The study was conducted in Mashuru and Makindu administrative divisions found in Kajiado and Makueni districts, south-eastern Kenya rangelands. The divisions occupy an area of about 3,100km² and 1,202 km² respectively. The study sites extend between latitudes 1°44’S to 2°26’S and longitudes 36° 50’E to 37° 45’E. The general altitude ranges between 1200m to 2000m above sea level. Among the most important features delineating the two divisions is the Nairobi- Mombasa railway line with Mashuru to the Northeast and Makindu to the West. This also marks the provincial boundary between the two divisions (Makueni district development plan 1997-2001, 1996 and Kajiado district development plan 1997-2001, 1996).

The Mashuru division is the ethnic homeland of the Maasai tribe, while Makindu is the ethnic homeland for the Kamba tribe as well as recent migrants from other tribes. According to the 1989 census, the Mashuru division had a population of 24,559 inhabitants with an average density of 6.8 persons per kilometre, characterised with higher population distribution around the water points and along the railway line and major roads (Waruru 2000). On the other, hand the Makindu division had a population of 34,522 inhabitants, with an average density of 29 persons per Km² (Makueni district development plan 1997-2001, 1996 and Kajiado district development plan 1997-2001, 1996).

The Mashuru and Makindu divisions are located in ecological zone V and VI, which are characterized by arid and semi-arid climatic conditions (de Leeuw 1991). These areas experience bimodal rainfall with a yearly mean of 600mm. The long rain comes in March to May, which leads to a long dry season lasting from June to September. The short rain is received from October to December, sometimes extending to January. A short spell of drought occurs in January to February. On average, Mashuru experiences a mean annual temperature range of 20.6°C to 23.0°C and Makindu 24.0°C to 21.6°C, while the potential evaporation for both divisions lies between 2000mm to 2400mm per annum (Shisanya 1996). Drought is a recurrent phenomenon in the climatic history of the study area, leading to high chances of crop failure (maize crop failure is to be expected during the short rain season every 1 in 3 years in the region) (Shisanya 1996). Vegetation of the study sites consists of trees, shrubs, herbs and grass which are adapted to the present climatic and ecological conditions. However, alteration of vegetation through climate, settlement, clearing and rangeland activities such as overgrazing and burning are common phenomena in much of the study areas (Waruru, 2000).

Methods

The study examined various extension communication methods employed to promote the adoption of weed management technologies to aid pasture improvement in the two communities of the south-eastern Kenya rangelands. Face to face interviews were conducted from August to November 2002. The sample unit employed was at household level with the head of the family being the main respondent. The sampling technique used was disproportionate and multistage, which involved a combination of stratified and systematic sampling techniques (Zikmund 2000). The area was stratified according to administrative divisions. To ensure appropriate representation was achieved, five locations in Makindu and four group ranches in Mashuru were identified for the study by extension agents and local leaders, as areas extensively encroached on by pasture weeds. The five administrative locations selected for sampling in Makindu division were; Makindu, Kalii, Kumoo, Kiboko and Kigauni. In Mashuru division, the four group ranches, which were currently under sub-division were: Masimba, Nkama, Kinyewa and Kima. Since grazing in Mashuru was not restricted to individual households, an assumption was made that communal ownership for grazing was still applicable.

The procedure for household selection included stratification of the divisions according to administrative location boundaries. With the assistance of both agricultural extension agents and local government administrators, a designated route was determined and, at an interval of 5km, two farmers on either side of the road were interviewed. In a situation where the designated farmer was unavailable, the neighbour was approached. The respondents were asked a number of questions about their perception and attitude towards the existing sources of information and communication channels in the promotion of pasture weed management technologies. In addition, they were also asked about the socio-economic benefits and values they attach to pasture weeds. A total of 165 sampling units; 83 and 82 from Makindu and Mashuru respectively were considered reasonably representative of the population, taking into account economic and time constraints.

Data analysis

An independent sample t-test technique was used to compare between the pastoral community (Mashuru division) and agro-pastoral community (Makindu division) regarding their perceptions and attitudes towards the existing communication channels and information sources used in extension of pasture weed management technologies. In addition, socio-economic values farmers in the two divisions attached to some pasture weeds were compared to establish if there was any relation with the adoption of pasture weed management technologies. Due to large sample sizes, the approximation to normality of means by the central limit theorem was applied and the Levene’s Test for equality of variances was conducted to test for homoscedasticity (Kenkel 1996). Equal variances were not assumed for F-statistic significant at less than 0.05 levels. The underlying principle was that if the two groups differed from each other regarding the perceptions and attitudes towards communication channels, sources of information and socio-economic values farmers attached to some pasture weeds, then a statistically significant difference in variance would be evident in the group means. A 95% confidence level (α =.05) was used in the calculations. Crosstabulation was used to supplement the t-tests in ascertaining the relationship between Makindu and Mashuru (Zikmund 2000). Interpretation of the crosstabulation was simplified by adopting a percentage calculation of relative variables and identifying the frequency of each dependent variable factor with each independent variable (division).

Results

Farm size and land tenure

The land use activities practiced in both divisions were a combination of livestock grazing and crop farming. Mashuru farmers had on average six times more cattle, two and half times more goats, five times more sheep and three times more donkeys than the Makindu farmers. Makindu farmers had on average three times the number of poultry and five times the number of beehives as the Mashuru farmers (Table 1).

Table 1: Average livestock herd size and beehive number owned by respondents according to divisions.

Livestock types and beehives	Study sites
	Makindu = 83		Mashuru = 82
	Mean (no.)	Range (no.)	Mean (no.)	Range (no.)
Cattle	8	0-22	47	0-140
Goat	28	0-85	74	0-255
Sheep	5	0-25	27	0-84
Donkey	1	0-3	3	0-10
Poultry	22	0-50	7	0-20
Beehives	5	0-15	1	0-10

In the Mashuru, an average land size of 1.7 ha was used for individual crop farming and an average of 21,637.3 ha was used for communal grazing where each farmer practiced weed control on an average of 0.7 ha. In the Makindu, the average individual farm size was 9.6 ha, of which 4.5 ha was used for crop farming and 5.1 ha was used for grazing. Of the total grazing land, a mean area of 1.3 ha per farmer was controlled for weeds (Table 2).

Table 2: Average farm size in hectares, land tenure and usage

Activities	Study sites
	Makindu n = 83		Mashuru n = 82
	Means(Ha)	Range (Ha)	Means (Ha)	Range (Ha)
Crop farming	4.5	1-14	1.7	0-6
Grazing land with weeds	5.1	0-15	21637.3	14873-39760
Weed managed area	1.3	0-5	0.7	0-7.6
Communal land tenure	-	-	21639	14877-39760
Freehold land tenure	9.6	1-23	-	-

Proximity of agricultural office and nearest town to farmer’s homestead

The proximity of farmers’ homesteads in the two divisions to the nearest agricultural office and shopping centre were compared. As shown in Table 3, there was a significant difference in the distance between the farmers’ homestead and the nearest agricultural office (P<0.0001) but no significant difference was found for the distance to the nearest shopping centre (P=0.288). Farmers in Mashuru were 24km away from government agricultural offices compared to farmers in the Makindu, who were 15 km away. It was found that some Mashuru farmers lived as far as 50km away from the agricultural offices while in the Makindu the greatest distance was only 38km (Table 4). The two divisions were almost equal distance from the nearest main shopping centre (8km for Mashuru and 6km for Makindu).

Table 3: Distance from agricultural office and nearest town to farmer’s homestead

Makindu n = 83 and Mashuru n = 82

Distance from	Division	Mean (Km)	SD	Range (Km)	t- value	P value
Agricultural office to farmer’s field	Makindu	15.42	8.55	3-38	-5.17	0.0001*
	Mashuru	24.66	13.75	2-50
Farmer’s homestead to the nearest town	Makindu	6	3.89	1-15	-1.07	0.288
	Mashuru	8	4.6	2-23

* indicates significant relationships between divisions with “t” probability < 0.05

Frequency of extension agents visits to farmers for providing advice on technology adoption

The frequency with which farmers in the two divisions were visited by various extension agents providing advice on pasture weed management technologies was compared. As shown in Table 4, a significant difference was found in the frequency of visits from: government extension agents (P=0. 002), opinion leaders (P=0.041 and non governmental organization extension agents (P=0.012). Government extension agents visited and provided advice to Makindu farmers on average three times per year (mean score 2.86) compared to the Mashuru farmers who were visited on average once per year (mean score 1.26). The Makindu community were visited by the Non Governmental Organisation (NGO) extension agent on average twice per year (mean score 2.7), compared to the Mashuru, who were visited on average once per year (mean score 1.36). Opinion leaders visited and provided advice to fellow farmers in Mashuru on average twice per year (mean score 1.93) compared to Makindu farmers, who were visited on average once per year (mean score 0.88).

As shown in Table 4, no significant difference in means was found between the two divisions in the frequency of visits by research scientists to farmers providing advice on pasture weed management (P = 0.061), though there was a tendency for Makindu farmers to be visited more frequently than the Mashuru farmers, probably due to the long distance between their farm and extension operation offices.

Table 4 Extension agents visits to farmers providing advice on weed management technologies

Makindu n = 83 and Mashuru n = 82

Attribute	Division	Significance rating values in percentage					Mean score	SD	t value	P value
		No visit	Yearly	Twice	Quarterly	Monthly
Government extension agent	Makindu	41	22	10	11	17	2.86	4.31	2.79	0.006
	Mashuru	60	24	5	5	6	1.26	2.91
NGO extension agent	Makindu	37	16	18	16	13	2.7	3.89	2.55	0.012
	Mashuru	56	24	7	6	6	1.36	2.92
Research scientist	Makindu	52	7	12	18	11	2.33	3.7	1.88	0.061
	Mashuru	61	18	7	7	6	1.35	2.94
Opinion leader	Makindu	72	18	4	1	5	0.88	2.60	-2.06	0.041
	Mashuru	56	23	5	4	12	1.93	3.87

Means were calculated by assigning values of responses 0 no visit; 1 for a visit yearly; 2 for visits twice per year; 4 for quarterly visit; and 12 for monthly visit
* indicates significant relationships between divisions with “t” probability < 0.05

Frequency of visits by farmers to extension offices in search of weed management technologies

A comparison was made between the two divisions in relation to the frequency of visits paid by farmers to extension agencies’ office to obtain pasture weed management information in cases where they were not visited by the extension agents. As shown in Table 5, a significant difference was found for visits paid by farmers to agricultural office (P<0.0001) and research stations (P<0.0001). It seemed farmers in Makindu visited the agricultural office and research institute for advice on pasture weed management at least twice per year compared to the Mashuru farmers who rarely visited. This was augmented by crosstabulation results, which indicated that about one third of farmers in Makindu visited agricultural office (34%) and research institute (37%) at least once a year, while the majority of Mashuru farmers (86.6%) rarely visited any of these offices.

The frequency with which farmers in both divisions visited various farms and training centres to obtain information on pasture weed management technologies was compared. As shown in Table 5, both t-test and crosstabulation showed no significant difference in the frequency of participation in farmer excursions (P=0.601), attendance at seminars/training/workshops (P=0.418), attendance at agricultural shows (P=0.27), visits paid to successful neighbour’s farm (P=0.805) and visits paid to opinion leader’s farms (P=0.324). This could be attributed to no participation in these activities by farmers in the two divisions as supported by the crosstabulation results.

Table 5 Farmers’ visit to extension information source

Makindu n = 83 and Mashuru n = 82

Attribute	Division	Significance rating values in percentage					Mean score	SD	t value	P value
		No visit	Yearly	Twice	Quarterly	Monthly
Agricultural office	Makindu	66	5	4	12	13	2.19	4.07	4.38	0.000
	Mashuru	87	6	7	0	0	0.2	0.56
Research station	Makindu	63	8	11	7	11	1.89	3.72	3.78	0.000
	Mashuru	83	11	2	4	0	0.3	0.84
Neighbours	Makindu	69	12	12	1	6	1.13	2.84	0.25	0.805
	Mashuru	77	10	0	8	5	1.02	0.00
Agricultural shows	Makindu	79	8	2	10	0	0.51	1.21	1.11	0.271
	Mashuru	91	2	6	0	0	0.29	1.40
Opinion leaders	Makindu	87	7	4	0	2	0.43	1.88	1.49	0.141
	Mashuru	96	1	0	2	0	0.11	0.63
Farmer excursion	Makindu	85	10	2	1	1	0.33	1.42	-0.52	0.601
	Mashuru	85	8	2	1	2	0.47	1.92
Seminar/ training	Makindu	96	1	1	1	0	0.08	0.49	0.81	0.418
	Mashuru	96	4	0	0	0	0.04	0.19

Means were calculated by assigning values of responses 0 for no visit; 1 for a visit yearly; 2 for visits twice per year; 4 for quarterly visit; and 12 for monthly visit
* indicates significant relationships between divisions with “t” probability < 0.05

Influence of extension agent’s cultural background in the adoption of weed management technology

The influence of cultural background of the extension agents in relation to that of the farmer in the dissemination of pasture weed management technologies was compared between the two divisions. As shown in Table 7, a significant difference was revealed for the cultural background difference between the extension agents and farmers in terms of: language (P=0.004), ethnicity (P=0.028) and gender (P=0.001).

The crosstabulation results supported the t-test by indicating that over 60% of farmers in the Mashuru considered language used by extension agents as a problem compared to about 40 % of the Makindu farmers (Table 6). On the other hand, no farmer (100%) in the Makindu considered the gender of extension agent as a barrier to extension of the pasture weed control compared to a few farmers (23%) in the Mashuru. The gender of extension agent therefore had influence to some Mashuru farmers in extension of pasture weed technologies.

No significant difference was detected between the divisions in relation to educational levels (P = 0.223), age (P = 0.215) and economic status (P = 0.320) of the extension agents and that of the farmers (Table 5). A majority of the respondents (83% for Makindu and 72% for Mashuru) in both divisions were not affected by the age of the extension agents, while almost 100% were not affected by education and economic differences (Table 5).

Table 6: Influence of background differences between farmers and extension agents

Makindu n = 83 and Mashuru n = 82

		Significance rating values in percentage
Attribute	Division	No barrier	Slight barrier	Moderate barrier	high barrier	Extreme barrier	Mean score	SD	t value	P value
Language barrier	Makindu	61	12	2	6	18	2.07	1.59	-2.93	0.004*
	Mashuru	37	12	10	16	26	2.82	1.66
Age difference	Makindu	83	0	0	0	5	1.63	1.41	-1.24	0.215
	Mashuru	72	2	2	8	15	1.91	1.56
Ethnic difference	Makindu	95	1	1	0	2	1.13	0.66	-2.20	0.028*
	Mashuru	85	1	2	4	7	1.46	1.19
Education difference	Makindu	100	0	0	0	0	1.00	0.00	-1.22	0.223
	Mashuru	98	1	0	0	1	1.06	0.45
Economic difference	Makindu	100	0	0	0	0	1.00	0.00	-1.00	0.320
	Mashuru	99	1	0	0	0	1.01	0.11
Gender difference	Makindu	100	0	0	0	0	1.00	0.00	-3.42	0.001*
	Mashuru	87	1	1	1	11	1.49	1.29

Means were calculated by assigning values of responses 0 no barrier, 1 slight barrier, 2 moderate barrier, 3 high barrier, and 4 extreme barrier.
* indicates significant relationships between divisions with “t” probability < 0.05

Influence of extension approach on adoption of weed management techniques

The influence of extension approaches used in the two divisions for pasture weed management technologies were compared. As indicated in Table 7, a significant difference between the divisions was found for the effectiveness of the participatory research approach (P=0.017). The Makindu farmers on average found the participatory research approach quite effective (mean score 3.2) compared to the Mashuru farmers, who found it only slightly effective (mean score 2.55). This was supported by crosstabulation results, which showed that about one half (53% )of the Makindu farmers considering participatory approach as either highly or extremely effective for dissemination of pasture weed management, compared to about one third (38%) of the Mashuru farmers.

There was no significant difference in the effectiveness of the on-research station approach (P=0.17). Farmers in both divisions considered on-station research approach as slightly effective means (mean score Makindu 2.04, Mashuru 1.73) of extension of pasture weed management technologies. These were supported in crosstabulation results, which showed about 70% of farmers in both divisions acknowledging that on-research approach was either not effective or only slightly effective in pasture weed control (Table 7).

Table 7 Extension approach commonly used in extension of weed management technologies

Makindu n = 83 and Mashuru n = 82

Attribute	Division	Significance rating values in percentage					Mean score	SD	t-value	P value
		Not effective	Slightly effective	Quite effective	Highly effective	Extremely effective
Participatory research	Makindu	32	6	8	14	39	3.20	1.74	2.40	0.017*
	Mashuru	51	6	5	12	26	2.55	1.76
On-farm research	Makindu	48	21	14	10	6	2.04	1.25	1.62	0.107
	Mashuru	63	15	9	8	4	1.73	1.15

Means were calculated by assigning values of responses 1 for not effective, 2 slightly effective, 3 quite effective, 4 highly effective and 5 extremely effective.
* indicates significant relationships between divisions with “t” probability < 0.05

Communication channels in extension of weed management technologies

The effectiveness of various communication channels in pasture weed management technology dissemination was compared between the two divisions. As shown in Table 8, a highly significant difference between the two divisions was found for the effectiveness of: radio (P<0.0001), televisions (P<0.0001), extension agents (P<0.0001), local organized groups (P<0.0001), demonstration plots (P<0.0001) and field days (P=0.027) in the dissemination of pasture weed management (Table 9). No significant difference between the divisions was found for the effectiveness of seminars (P=0.056), pamphlets (P=0.088), brochures (P=0.358) and newspapers (P=0.16 in dissemination of pasture weed technologies.

As shown in Table 8, the Makindu farmers on average found the use of radio, televisions, demonstration plots, organized groups and field days as slightly effective media for acquiring information related to pasture weed management compared to the Mashuru farmers who found the same media virtually ineffective. Farmers in both divisions considered information sourced from newspapers as slightly effective but information sourced from pamphlets, seminars and brochures was virtually ineffective. On inspection of the distribution of crosstabulations, it supported the t-test results (Table 9)

Table 9 Information dispersal channels in adoption of weed management technologies

Makindu n = 83 and Mashuru n = 82

Attribute	Division	Significance rating values in percentage					Mean score	SD	t-value	P-value
		Not effective	Slightly effective	Quite effective	Highly effective	Extremely effective
Radio	Makindu	32	16	12	16	24	2.83	1.61	5.13	0.0001*
	Mashuru	68	12	7	6	6	1.70	1.21
Television	Makindu	60	7	10	7	16	2.11	1.55	4.40	0.0001*
	Mashuru	91	2	0	2	4	1.24	0.88
Extension agents	Makindu	57	11	13	4	16	2.11	1.51	3.82	0.0001*
	Mashuru	83	6	5	4	2	1.37	0.92
Organized groups	Makindu	58	12	7	10	13	2.08	1.50	5.48	0.0001*
	Mashuru	93	2	2	0	1	1.12	0.55
Demonstration	Makindu	60	5	5	12	18	2.23	1.66	3.33	0.001*
	Mashuru	79	6	4	7	4	1.50	1.10
Field days	Makindu	76	5	7	12	0	1.75	1.45	2.24	0.027*
	Mashuru	89	2	5	4	0	1.32	0.98
Seminar	Makindu	94	1	1	2	1	1.16	0.67	1.92	0.056
	Mashuru	99	1	0	0	0	1.01	0.11
Pamphlets	Makindu	82	6	4	5	4	1.42	1.03	1.72	0.088
	Mashuru	93	2	1	1	2	1.18	0.74
Brochure	Makindu	82	5	4	4	6	1.47	1.13	0.923	0.358
	Mashuru	89	2	1	2	5	1.32	0.99
Newspapers	Makindu	72	8	4	5	11	1.73	1.37	1.41	0.160
	Mashuru	85	4	1	0	10	1.45	1.21

Means were calculated by assigning values of responses of 1 not effective, 2 slightly effective, 3 quiet effective, 4 highly effective and 5 extremely effective.
* indicates significant relationships between divisions with “t” probability < 0.05

Perceived important values respondents attached to some pasture weeds

The values which farmers in the two divisions attached to some pasture weeds were compared. As shown in Table 9, no significant difference was found for the importance of some pasture weeds to farmers as: livestock forage ( P=0.633) and firewood (P=0.053), although there was a tendency for the Makindu farmers to value pasture weeds more as a source of firewood. In essence, pasture weeds were moderately valued in both divisions as a source of livestock feed and firewood, which was supported in crosstabulations.

As shown in Table 9, a highly significant difference between the divisions was found on the values the farmers attached to pasture weeds as: medicinal plants (P<0.0001), construction material (P=0.007), and wild food reserve (P<0.0001). The Mashuru farmers on average placed more importance on some pasture weeds as a source of medicinal plants and construction material than the Makindu farmers. Of note, most farmers in the Mashuru (86%) believed that some pasture weeds were either a highly or extremely important source of medicinal plants compared to only a small group of Makindu farmers (15%). In addition, the majority of farmers (93.9%) in the Mashuru did not value pasture weeds as a source of wild food compared to 59% of the Makindu farmers.

Table 10 Values respondents attached to some pasture weeds

Attributes	Division	Significance rating values in percentage					Mean score	SD	t-value	P-value
		No value	Little value	Moderate value	high value	Extreme value
Forage	Makindu	14	29	6	18	32	3.25	1.52	-0.45	0.633
	Mashuru	17	15	18	15	35	3.37	1.51
Firewood	Makindu	12	12	23	23	31	3.49	1.36	1.95	0.053
	Mashuru	17	22	19	19	22	3.07	1.41
Medicinal value	Makindu	54	18	12	7	8	1.98	1.32	12.74	0.0001*
	Mashuru	4	5	6	26	60	4.33	1.04
Construction material	Makindu	18	19	18	32	12	3.01	1.32	-2.74	0.007*
	Mashuru	11	15	12	29	33	3.59	1.37
Wild Food	Makindu	59	25	7	6	2	1.67	1.01	3.96	0.000*
	Mashuru	94	2	1	0	2	1.15	0.67

Means were calculated by assigning values of responses 1 no value, 2 little value, 3 moderate value, 4 high value and 5 extreme value
* indicates significant relationships between divisions with “t” probability < 0.05

Discussions and recommendations

The role of extension personnel in the transfer of agricultural technology has been found to be crucial in the achieving of high adoption levels (Guerin 1994). However, prior to technology dissemination, extension agencies need to explore and understand farmers’ needs and the relevance of technology for that particular situation. This will in turn influence the kind of extension approach and technology offered to farmers for adoption. The process of gathering information and enhancing adoption is directly linked with the frequency of interaction between farmers and extension agents (Rogers 1995 and Crouch et al. 1981). For instance, infrequent visits and provision of advice to farmers by extension agents leads to limited awareness of appropriate technologies and consequent non adoption of technologies.

Although the adoption of pasture weed management technologies in both divisions was generally low, the frequency of visits made by government extension agents and the reciprocal visits paid by farmers in the Makindu division was more than in the Mashuru division (Table 4&5). This infrequent interaction between the farmers and government extension agents in the Mashuru appeared to have contributed more significantly to the limited knowledge of pasture weed management, which is represented from the lesser area controlled of weeds (Table 2). On the other hand, the Makindu farmers were more frequently visited and reciprocated and therefore were more aware of pasture weed management technologies, with over a greater area controlled of weeds using the recommended methods. Lacefield et al. (1998) noted that human influence and face-to-face interaction were important factors in the technology transfer and adoption process. Since the Makindu divisions were visited by government extension agents at least twice a year, they were more aware of the importance of weed clearing and the appropriate techniques compared to the Mashuru community. However, more impact of adoption in both divisions would have been felt if the frequency of contact between the extension agents and farmers would have been increased. This was supported by a case of two communities in Kenya with regard to the rate of adoption of agricultural production (Bohlin 2002). The Muchohi/Uruma group which was regularly visited by government extension agent had a better agricultural production than the Karindundu group, who were never visited.

Contrary to developing nations, the emerging sophisticated electronic communication in Australian agriculture has made the sourcing of information less significant as a constraint to adoption (Guerin et al. 1994). However, there are cases in Australia to compare with the finding of this study. For instance, the frequency of visits paid by government extension agent and the rate of adoption found in the two divisions are comparable with the two groups of farmers in north central Victoria, Australia, concerning the adoption of Lucerne farming. The group that had been frequently visited by government extension agents adopted Lucerne farming while the group that had no visits did not adopt the technology since they were not aware of the technology (McCarthy et al. 1997).

According to Sunding et al. (2000), geographical location and distance to regional centres with information and services influence the level of adoption. Thus, producers in locations farther from a regional centre where most services are available adopt later since they are affected by travelling costs and time. This phenomenon of remoteness and proximity of farms to agricultural offices and research institutes seemed apparent in the current study. The average distance between the agricultural office and farmers in the Mashuru division was found to be farther than in the Makindu (Table 3). As a result, limited visits were made by extension agents to farmers’ fields in the Mashuru, unlike in the Makindu, which received more contacts, leading to a relatively greater awareness of recommended weed management technologies. This interpretation of the finding is consistent with the findings of the Australian Department of Agriculture, Fisheries and Forestry (1999) that isolation and remoteness of Western and Northern Australia to most pastoral resource stations and neighbours was a serious constraint to the implementation of landcare practices.

Other major sources of knowledge of new practices recognized in developing countries are neighbours, family and community based organizations. It appeared from the study that the vastness of Mashuru division limited the interaction among farmers and successful neighbours in sharing of information about new technologies. The Makindu division had a comparative advantage due to closer linkages that was evidenced from smaller farms sizes (Table 2). This ‘closeness’ was thought to have also contributed to a relatively higher adoption rates. According to (Rogers 1995) a ‘chain reaction’ occurs in communities that are closely linked but not in those that are sparsely distributed. Foster (1995) found successful neighbours to be influential in extension of new technologies in a social system. In portraying the importance of successful neighbours, Hassan et al. (1998) demonstrated that an increased number of farmers using new variety maize significantly influenced others to adopt. Another probable explanation, especially for the Mashuru division, could be long distances between the neighbors and heterogeneity in culture as a result of migrants from different cultural background settling in the area, leading to limited information sharing (Rogers 1995).

In addition to government extension agents, several organizations and actors are involved in disseminating of agricultural knowledge and skills in different parts of Kenya (Muok et al. 2001). The intention is to create more awareness of technologies to farmers, which has a corresponding relationship to levels of adoption. In other words, the higher the extension input to farmers the higher the level of adoption (Floyd et al. 2003). The study revealed that the Non Governmental Organization (NGO) agents visited farmers in both divisions at least once per year (Table 4). However, in many circumstances, the activities of NGOs in the regions were related to soil and water conservation, tree planting and other agricultural activities (Kaluli 1993), and less consideration was given to livestock and pasture improvement strategies. If, however, the frequency of NGO’s visits could be increased and include in their goals specifically the extension of pasture weed management technologies and other necessary support provided, then pasture improvement through weed control would probably be enhanced in the region.

Opinion leaders are informally influential people in a social system and they act as entry point for technological dissemination (Rogers 1995). Innovations approved by opinion leaders in a social system are more likely to be adopted by followers irrespective of their risks and uncertainties (Guerin 1994). The study found opinion leaders more frequently visited and provided advice to fellow farmers in the Mashuru division than in the Makindu division (Table 5). However, their contribution to pasture improvement through weed control was minimal. This was evidenced from the low adoption of weed management technologies in Mashuru, as the areas controlled of weeds were small relative to the grazing land covered by weeds (Table 2). This study suggested that opinion leaders were also limited in knowledge of weed management technologies and had no resources or materials to aid in the dissemination of information. To enhance the effectiveness of opinion leaders in pasture weed management, there is a need to equip them with technically accurate information and in a form that can be passed on.

Personal background of extension agents in weed control

While ethnicity appears to have an impact on intra-community communication, it may also be an impediment where the cultural background of the change agents (extension officer) is dissimilar to the one of potential adopters (Rogers 1995). This observation was consistent with the finding of this study that the language barrier between the extension agents and pastoralists in Mashuru was considered an issue (Table 6). Communicating with farmers in the local language was observed to be more effective than using foreign languages such as Swahili or English or an interpreter, as is common with extension services in Kenya.

Considerable attention has been given in recent years to the fact that a large proportion of agricultural work in many countries is done by women, whereas only a small proportion of the agricultural extension agents are female (Ban et al. 1988 ). However, this disparity in proportion of male to female extension agents is necessary and should differ among communities based on different attitudes and beliefs about the gender of extension agent in cultural context. For instance, gender difference was perceived as a greater issue in extension service for communities in Mashuru than in Makindu (Table 6). It is unacceptable to the Maasai culture (predominantly in Mashuru) to give leadership to females or children; this is still the domain of the patriarch (Aboud 1986). This might have hampered and probably contributed to limited extension services in pasture weed control in Mashuru as compared with their counterparts in Makindu.

Economic and education level difference were not reported as constraints to extension in this study. The respondents believed that as a long as the extension agents are skillful enough to impart appropriate knowledge, their level of education and economic status did not matter. However, there was a slight perception among respondents in the Mashuru that extension agents of their own tribe were more appropriate and effective than those from other tribes even when they have knowledge of the local language.

Influence of technology transfer approach in pasture weed control

There are various factors that extension agents should consider when choosing an extension approach. These include the cultural set up of the target community, education level of the target group and above all the manpower and resources available (Ban et al. 1988). The finding that participatory research was more used by extension agents in the Makindu division than in the Mashuru (Table 7) was consistent with Franzel (2002) findings on the adoption of improved fallows to curb soil fertility in Kenya. In the participatory research, farmers’ problems are collaboratively diagnosed and incorporated into the research development, while information dissemination is targeted to reflect cultural views (Franzel 2002 and Aboud 1986). Participatory research also involves higher levels of ‘face to face’ contact, which was reflected in the frequent visits made by extension agents and research scientists to Makindu farmers, a factor lacking in contact with the Mashuru and accounting for their lower levels of weed management adoption.

This finding of minimal use of participatory research approach in Mashuru was in line with the findings of Supriadi et al. (1999) in relation to the adoption of rubber research technologies in Indonesia, where most research priorities were predetermined by researchers and were mainly conducted on research stations and transferred to farmers for adoption through extension agents. This implied that farmers were considered as recipients of technology and extension agents as merely messengers without any feedback. To improve the adoption in both divisions, the study suggested that weed management technology testing needs to be participative as this would identify client’s needs, problems, values of weeds and resource endowment.

Effectiveness of mass media and inter-personal channels in promotion of pasture weed management technologies

Passive audiences have to be lured, so the message needs to be highly visual, using catchy themes and examples of successes demonstrating ease of achievement and an appeal to self interest (Lloyd, 2003). This is also consistent with the promotion and marketing of various pasture weed management technologies for adoption in arid and semi arid lands. The study clearly demonstrated how various information dispersal channels influenced the adoption of weed management technologies in south-eastern Kenya rangelands. The lack of demonstration sites and field days were identified as a major constraint to the adoption of pasture weed management technologies by farmers in this study (Table 7). Other African studies (Marsh 1998) noted that farmers preferred demonstration plots and field days for the improvement of agricultural development over other communication means. Thus, through demonstrations farmers are convinced about the appropriateness and workability of a technology, which lead to reduced uncertainties and risks surrounding the adoption process.

The poor telecommunication infrastructure in remote regions of Mashuru and to a lesser extent the Makindu exacerbates the already substantial isolation factors experienced by people living in rangelands, which include long distances to services, poor road network and lack of telecommunications. In this study, there was a significant difference in the way each region employed radio, television, newspaper, pamphlets and brochures in promoting pasture weed management programs (Table 8), though overall use of all these types of media was minimal. Radio was rated first in Makindu and second in Mashuru, however, the perception scale showed as a medium it was minimally effective in both divisions (Table 8). It appeared that weed control content in broadcasts was limited, hence the limited knowledge among farmers in the study regions. Language used could also account for the poor credibility of radio as broadcasts are in English or Swahili, different languages to the one commonly understood by the respondents in the two divisions. Lingamneni (1981) observed that the type of language and relevance of the messages aired on media channels had an implication on the adoption of new technologies in developing nations. Among the Nigerian farmers, the use of radio in extension was rated second with a higher perception scale (Means scale 4.11) while extension agents took first position (Ekoja 2002). This study’s indication of low perception of the effectiveness of radio reflects poor utilization of this medium by government agencies and in particular by extension staff.

In both communities there was limited attendance by farmers to agricultural shows, “farmer-exchange” visits and seminars/trainings related to weed management technologies (Table 8). Similar to the argument advanced for the field days, this accentuated the inadequate exposure of farmers to weed management technologies across the study sites. Various studies have showed that attendance at agricultural shows, field days and seminars increased awareness and knowledge about the new practices and consequent adoption (Rogers, 1995).Through farmer-exchange visits, Kiara (1997) observed that farmers were able to learn from fellow farmers in the implementation of soil conservation measures. In situations without such opportunities, the adoption rate is minimal as was demonstrated by the two communities in this study.

Perceived values respondents attached to pasture weeds

Although pasture weeds were considered problematic to livestock productivity, some farmers tended to place value on some weeds as a source for their socio-economic needs, which in turn also contributed significantly to the existence and non adoption of pasture weed management in south-eastern Kenya rangelands. For instance a livestock keeper with a large herd of browsers prefers to keep a large proportion of palatable woody weeds while grazers encourage herbaceous forage and possibly keep the woody plants mainly for shade. Additionally, in the long dry season, the commonly available and dependable feeds for livestock are woody weeds due to their drought tolerance characteristics.

The role of fire wood in arid and semi-arid lands as source of fuelwood can not be overemphasized, and the woody weeds are an integral part of these sources (Kirubi et al. 2000. The findings of the study indicated that the two communities rationally encouraged the encroachment of some pasture weeds because they were sources of fire wood. This value of fire wood from woody weeds reflected the fact that 95-98% of the national fuelwood demand and the total energy supply in Kenya were sourced from woody vegetation (Kirubi et al. 2000). According to Brown et al. (1999), dependency on fuelwood in developed nations declined with the discovery of alternative sources of energy such as oil, coal and solar.

In developing countries, some woody plants provide raw materials for construction of huts, fences, granaries (traditional stores), kraals, hanging of beehives and beehives, thatching, weapons, woodcarvings, musical instruments and clothing (Burley (1982). The building styles of the houses reflect the cultural diversity and potential needs of raw material for construction. For instance, most the houses in Mashuru (locally referred to as “Manyatta”) are temporary constructions of a mixture of dung and mad smeared on sticks and wooden poles, sourced from the surrounding vegetation (Coast 2002). Under such circumstances, pastoralists act unfavorably towards practices that do not conserve useful resources such as woody plants for construction purposes, despite the effects they caused on natural pastures.

As an important source of medicinal plants, communities in Mashuru placed far more importance on some woody weeds than their counterparts in Makindu. The difference in valuing medicinal plants between the two communities seemed associated with the knowledge about the plants, which has religious connotations among most African communities (Rashid 2000). Furthermore, the use of herbal medicines among the Maasai is of major importance in treating various diseases and it is considered the responsibility of every member of the household to know them and their uses (Sindiga 1994). Given the importance placed on woody weeds by the Maasai community and the Mashuru in particular, it was most unlikely for them to adopt the management approaches that aimed at removing medicinal plants despite their detrimental characteristics on pasture. From this finding, we suggest further research to identify and register important medicinal plants in the region, particularly in Mashuru where there still a large body of indigenous knowledge about medicinal plants. We recommend that after identification of these plants, every effort should be made for their conservation and utilization along with conventional medicines.

There was some importance placed on some natural pasture weeds as a source of food in the Makindu divisions. This was not true in the Mashuru divisions (Table 5). However, the overall scenario indicated minimal perception of pasture weeds as source of food across the divisions despite the significant difference detected in the test. The results indicated that little attention was given to pasture weeds by the two communities for the supply of wild fruits and other wild foods in favour of pasture improvement. Whereas wild fruits were important source of food in rangelands, Muok (2000) and Brown (1999) found exotic and domesticated wild fruits increasingly becoming dominant in this area. This in turn explained why the two communities placed less importance on pasture weeds with fruit value in preference for pasture improvement.

Conclusion and implications

In examining community responses to the various communications media, it is concluded that most were ineffectively utilized for dissemination of weed management technologies, so leading to non adoption of weed management technologies in the study sites. Paradoxically, the study found farmers placing strong value on some pasture weeds, which in essence contributed to less adoption of weed management technologies. The processes promoted by a strong ‘top-down’ approach to management on the part of government extension agencies and infrequent interaction between extension agents and farmers failed to identify farmers needs and values towards resources leading to wrong assumptions made against farmers. Communities have their own culture, aspirations and value systems, which must be recognized when involving them in natural resource management. The process of planning and consultation is as important as the outcome and should be aimed directly at the needs of the affected community. Extension officers are often key individuals in the design of communication strategies and maintaining face to face contact with communities. This study shows a need to strengthen and better integrate communication, education and community involvement when formulating and implementing weed management strategies. In addition, the study has shown that remoteness and isolation of communities limit interactions, leading to a lack of sharing of technical information and consequently hampers uptake of development practices such as pasture weed management.

Reference

Aboud AA (1986). Range development constraints and research in Kenya. In ‘Range development and research in Kenya’ (Eds. R.M. Hansen.; B.M. Woe.; R.D. Child). Proceedings of a conference, Agricultural Resource Centre, Egerton University, Njoro, Kenya

Ban van de AW and Hawkins HS (1988). Agricultural extension, Longman house, burnt mill, Harlow

Barrow EGC (1991). Evaluating the effectiveness of participatory agro-forestry extension programmes in a pastoral system, based on existing traditional values. A case study of Turkana in Kenya. Agro-forestry systems 14,1, 1-21.

Bohlin AGM (2002). Do agricultural extension services really matter? A study on the significance of agricultural extension with special focus on women farmers in Mathira division of Kenya. Minor field studies- International Office, Swedish University of agricultural sciences, 217, 78

Brown R.; Flavin, C.; French, H.; (1999), The state of the world; A worldwatch institute Report on progress toward a sustainable society, Linder Starke, series editor, W???)

Burley, J (1982), Obstacles to tree planting in arid and semi arid land: Comparative studies from India and Kenya, (online), http://www.unu.edu/unupress/unupbooks/80391e/80391E07.htm [accessed 3^rd April 2002]

Crouch BR.; Chamala S (1981). Communication strategies for technological change in agriculture: Implication for rural society. Extension, eduction and rural development 267-275

Chamala S.; Coughenour MC (1987b). Model for innovation development, diffusion & adoption (MIDDA): A system for agricultural development. Journal of extension systems, Vol. 3 no. 1 pp. 47-55.

Coast, E. (2002), Maasai socio-economic conditions: a cross-border comparison. Human ecology, Vol. 30, no. 1 pp. 79-105

Department of Agriculture, Fisheries and Forestry-Australia 1999, Implementing landcare in the rangelands. The Pastoral and Graziers Association (PGA) of Western Australia (Inc.)

Ekoja II (2002). Channels of communicating agricultural information preferred by Nigerian farmers. Quarterly Bulletin of IAALD, Vol. XLVII, 2, 38-42

Floyd C.; Harding AH.; Paudel KC.; Rasali DP.; Subedi K.; Subedi PP (2003). Household adoption and the associated impact of multiple agricultural technologies in the western hills of Nepal. Agricultural systems 76, 715-738

Foster AD.; Rosenzweig MR (1995). Learning by doing and learning from others. Human capital and technological change in agriculture. Journal of political economy 103, 6.

Franzel S.; Phiri D.; Kwesiga F (2002). Assessing the adoption of improved fallows in Eastern Zambia. Trees on farm: assessing the adoption potential of Agroforestry practices in Africa, 37-64

Guerin^A LJ.; Guerin TF^BC (1994). Constraints to the adoption of innovations in agricultural research and environmental management: a review. Australian Journal of experimental agriculture 34, 549-71

Hassan RM.; Karanja DD.; Mulamula HHA (1998). Availability of effectiveness of agricultural extension services for maize farmers in Kenya. In ‘Maize technology development and transfer’ (Ed. Hassan R.M) a GIS application for research planning in Kenya, pp. 175-188, CAB international, Wallingford, UK

Kenkel JL (1996). Introductory statistics for management and economics, Duxbury press, Belmont

Kiara JK.; Skoglund E.; Eriksson A (1997). Development of soil conservation extension in Kenya. Soil conservation extension: from concepts to adoption. In (Eds. Sombatpanit, S.; Zobisch, M.A.; Sanders, D.W.; Cook, M.G. pp. 35-45, Enfield, USA

Kajiado District development plan 1997-2001, (1996) Office of the president and ministry of planing and national development, Nairobi, Kenya.

Kaluli J (1993). Involvement in agricultural activities in Machakos district. Non-governmental organizations and the state in Africa: In ‘rethinking roles in sustainable agriculture development’ (Eds. Wellard, K.; Copestake, J.G), 111-117

Kirubi, C.; Wamicha, WN.; Laichena, JK (2000). The effects of woodfuel consumption in the ASAL areas of Kenya: the case study of Marsabit forest. African Journal of ecology Vol. 38, no. 1, pp. 47-52

Lacefield GD.; Ball DM.; Gerish JR.; Hannaway DB (1998). Technology transfer- perception of extension agronomists. Proceedings of 9^th Australian agronomy conference, Wagga wagga. www.regional.org.au/au/asa/1998/plenary/lacefield.htm

de Leeuw PN (1991), Introduction to the Kenyan rangelands and Kajiado district. In (Eds. Solomon, Bekure.; P. N. de Leeuw.; B.E. Grandin). An analysis of the livestock production system of Maasai pastoralists in eastern Kajiado district, Kenya. International Livestock Centre for Africa, Addis Ababa, Ethiopia

Lingamneni JR (1981). Communication channels/Sources in the innovation-decision process: A Comparative Review of U.S and India diffusion research. In Crouch, B.R and Chamala. S, (Eds). Extension education and rural development.1, 147-153

Lloyd D 2003. Communication course notes, Science for management, Southern Cross University, Australia

Makueni District development plan 1997-2001 (1996). Office of the president and ministry of planing and national development, Nairrobi, Kenya.

Muli JM.; Omanga PA.; Jones RB (1997). Farmer-to-farmer seed supply: case study pigeonpea seed distribution in Kenya. Alternative strategies for smallholder seed supply. Proceedings of an international conference on options for strengthening national and regional seed systems in Africa and West Asia, Harare, Zimbabwe, pp. 188-191

Muok B.; Kimondo J.; and Atsushi L 2001. Farmer to farmer extension: experience in drylands Kenya Proceedings of the extension working party (S6.06-03) symposium 2001. Assisting forest owner, farmer and stakeholder decision-making International Union of Forestry Research Organizations. The regional Institute Ltd. [Online] www.regional.org.au/au/iufro/2001/muok.htm [accessed 30th September 2001]

Maranga EK (1998). A review of range production and management extension activities in Kenya, Journal of agricultural and environmental ethics, Vol. 11, no. 2, pp. 131-144

Marra M (2001). The economics of risks, uncertainty and learning in the adoption of new agricultural technologies: Where are we on the learning curve? www.general.uwa.edu.au/udpannel/dpap0110.htm [accessed 5^th April 2002]

Marsh SP.; Pannell DJ (1998). What we think we know about extension and why its not enough for landcare, SEA Working paper 97/03, [online], www.general.uwa.edu.au/u/dpannell/spap973f.htm, [accessed 5^th April 2002]

McCarthy MP.; Oxley T (1997). Farmer adoption of dry land Lucerne systems in North Central Victoria , Australia

Pontius SK (1983). The communication process of adoption: Agriculture in Thailand, Journal of developing areas, Vol. 18, pp. 93-118.

Rashid, AZMM.; Khatun, M. (2000). Medicinal plants- a potential agroforestry component in Bangladesh: agroforestry today Vol.12, no. 1 (eds. Dali Mwagore & Helen van Houten)

Rogers EM (1995). Diffusion of innovations, 4^th Edition, the Free press, a division of Macmillan, inc.

Shisanya CA (1996). Chances and risks of maize and bean growing in the semi arid areas of South east Kenya during expected deficient, Normal and above normal rainfall of the short rainy seasons. Materialen Zur Ostafrika- forschung, Heft 14. GGT, Trier

Sindiga, I. (1994). Indigenous (Medical) knowledge of the Maasai. Indigenous knowledge & development monitor, Vol. 2, no. 1 pp. 16-18

Sunding D.; Zilberman D (2000). The agricultural innovation process: Research and technology adoption in a changing agricultural sector (For handbook of agricultural economics)

Supriadi M.; Chamala S (1999). Challenges in implementing participatory research approach: A case of Indonesian rubber industry. The journal of agricultural education and extension, 6, 2

Waruru BK (2000). The soils of Mashuru Division, Kajiado District: Information to improve soil productivity in arid environment. Reconnaissance Soil Survey report no. R18, 2000.

Zikmund WG (2000). Business Research methods, 6^th ed., Dryden press USA