Funded by the Maize and Wheat
Euroandd U o Systems for the Poor
Adoption of Maize
Production Technologies in
1 Selian Agricultural Research Institute, Arusha, Tanzania.
2 CIMMYT, Economics Program, Addis Ababa, Ethiopia.
3 SACCAR, Gaborone, Botswana.
4 Ilonga Agricultural Research and Training Institute, Kilosa, Tanzania.
CIMMYT is an internationally funded, nonprofit scientific research and training organization. Headquartered in
Mexico, the Center works with agricultural research institutions worldwide to improve the productivity and
sustainability of maize and wheat systems for poor farmers in developing countries. It is one of 16 similar centers
supported by the Consultative Group on International Agricultural Research (CGIAR). The CGIAR comprises
over 50 partner countries, international and regional organizations, and private foundations. It is co-sponsored by
the Food and Agriculture Organization (FAO) of the United Nations, the International Bank for Reconstruction
and Development (World Bank), the United Nations Development Programme (UNDP), and the United Nations
Environment Programme (UNEP).
Financial support for CIMMYT's research agenda currently comes from many sources, including the governments
of Australia, Austria, Belgium, Canada, China, Denmark, France, Germany, India, Iran, Italy, Japan, the Republic
of Korea, Mexico, the Netherlands, Norway, the Philippines, Spain, Switzerland, the United Kingdom, and the
USA, and from the European Union, the Ford Foundation, the Inter-American Development Bank, the Kellogg
Foundation, the OPEC Fund for International Development, the Rockefeller Foundation, the Sasakawa Africa
Association, UNDP, and the World Bank.
Printed in Mexico.
The impact of maize research and extension in Tanzania's Northern Zone during the past 20 years was evaluated
through a formal survey in 1995. Data were grouped into two major agroecological zones: the intermediate and
the lowland sub-zones. Sample farmers were also categorized based on land preparation method. A two-step
Heckman's procedure was used to simultaneously analyze factors affecting adoption of improved maize seed and
inorganic fertilizer. The study found that demand for composite seed was less than that for hybrids, although the
National Maize Research Program had released more composites, which can be recycled. Farming experience
was the only factor that significantly influenced the probability of adopting improved maize in the intermediate
zone. No factor significantly influenced intensity of adoption of improved seed. About 80% of farmers recycled
improved varieties, including hybrids, contrary to recommendations. The rate of adoption of chemical fertilizers
was low, influenced only by the number of livestock units. No farm characteristic influenced intensity of fertilizer
adoption. Recommendations on fertilizer placement were poorly followed, which i, I-inf -, the negative impact
of the poor management of crop residues in the zone. Formal credit is not available to maize farmers; with rising
input prices, this dynamic will become more critical. Adoption of recommendations on land preparation,
frequency and timing of weeding and fertilizer application, and plant spacing have been successfully adopted in
both zones. Recommendations based on survey results include the development of additional hybrids for the
Northern Zone and/or village level production of composite seed; research on the economics of recycling
improved varieties (including both composites and hybrids); more research and extension effort directed toward
efficient use of fertilizers (manure, chemical fertilizer, and crop residues); and encouraging measures by banks and
policy makers to make credit more available to small maize farmers with a high rate of loan recovery and low cost
Correct citation: Nkonya, E., P. Xavery, H. Akonaay, W. Mwangi, P. Anandajayasekeram, H. Verkuijl, D.
Martella, and A. Moshi. 1998. Adoption of Maize Production Technologies in Northern Tanzania. Mexico
D.F: International Maize and Wheat Improvement Center (CIMMYT), The United Republic of Tanzania, and the
Southern African Center for Cooperation in Agricultural Research (SACCAR).
AGROVOC descriptors: Tanzania; Maize; Zea mays; Varieties; Hybrids; Plant production; Socioeconomic
environment; Credit policies; Production factors; Crop management; Cropping patterns; Crop residues; Fertilizer
application; Weeding T. 1 i. 1. .. transfer; Innovation adoption
AGRIS category codes: E14 Development Economics and Policies
E16 Production Economics
Dewey decimal classification: 338.163
T ab les ....................................................................................................................... iv
Figures ................................................................................................................... v
Abbreviations and Acronyms......... .......... .............................................................. ...... vi
Acknowledgments .......................................................................................................... vii
Executive Summary ....................................................................................................... viii
1.0 Introduction ................................................................................................... 1
1.1 Introduction to the study .................. .. .. .... ... ... .. ............ ................... 1
1.2 Objectives of the study ........................................................................ ............ 1
1.3 Description of the study area ................................................................... ............... 2
1.4 Methodology ................................... ................. ............ 4
2.0 History and development of maize research............................................................ 7
2.1 Introduction ..................................................... ........... 7
2.2 Maize production t. I. 1111 -.- recommendations ........... ............................. ............... 8
3.0 Socioeconomic and demographic characteristics ..................................... ......... 12
3.1 Demographic characteristics .................................................................................. 12
3.2 Land resources and allocation patterns...................................................... ................... 12
3.3 Livestock ownership ...................... .... .... ... ... .. ........ ........................ 13
3.4 Farm m echanization ....................................................... ................. ................ 14
4.0 Maize production practices and adoption of recommendations ............................... 16
4.1 Crops and cropping systems .......................... .................................................................. 16
4.2 Cropping calendar ................................................. .......... 16
4.3 Rate and intensity of adoption of improved maize varieties ............................................... 18
4.4 Factors affecting adoption of improved maize varieties ................................................. 19
4.5 Seedbed type, planting configuration and weeding ............................................. .......... 26
4.6 Fertility management ...................................................... .................. ........ .. 28
4.7 Pest and disease control .................................................... .................................... 32
4.8 Transportation, storage, and post-harvest t., 1, in. .1 ,......................................................... 33
4.9 Seed selection .................................... ...... ..... .. ............. 34
5.0 Credit and extension services ............................................................................ 36
5.1 Credit availability ..................................................... ........................... 36
5 .2 S sources of inform action ............................................................... ................................ 3 7
6.0 Conclusions .......................................................................................................... 39
References ............................................................................................................... 41
1. Commercial varieties in Tanzania and their potential........................................... ................. 9
2. Fertilizer recommendation for maize according to agroecological zones ....................................... 10
3. Demographic characteristics of sample households ....................................................... 12
4a. Livestock ow nership on zonal basis ................................................. ................................ 14
4b. Livestock herd size by land preparation .......................... .................................. ........................ 14
5. Num ber of farm im plem ents owned ................................................ ................................ 14
6. M ethod of land preparation in each zone .............................................. ........................... 14
7. Cropping system s in the tw o zones ................................................. ................................ 16
8. Heckman's first stage procedure results estimating factors affecting adoption of
improved maize seed for the intermediate and lowland zones ................................................. 20
9. Heckman's first stage procedure results estimating factors affecting adoption of
improved maize seed and fertilizer for the two zones ..................................................... 21
10. Heckman's second stage procedure SUR results estimating factors affecting intensity of
adoption of improved maize seed and fertilizer for intermediate zone and for the two zones.......... 22
11. Trend of maize produce price to input price ratio in Tanzania................... ................................... 24
12. Maize varieties planted in the 1994/95 season ....................................... .................. 25
13. M ost preferred m aize varieties .................................................... .................................... 25
14. Reasons for varietal preference ................................................... ............................... ..... 25
15. D discontinued varieties .................... ............. ..... ..... .... ... ....... ........... ..... 26
16. Percentage of farmers planting on flat beds or ridges in each zone ............................................ 26
17. Spacing for monocropped maize and maize/legume intercrop...................................................... 27
18. Frequency of w feeding ...... ............... ............ .... .... .... .... ......... ........... ..... 27
19. Time of planting, weeding, and fertilizer application ....................................... ............... 27
20. Adoption of fertilizer use in the 1994/95 season ....................................... .. ............... 28
2 1. Fallow ing and crop rotation ...................................................... ...................................... 29
22. Trend of fertilizer application ..................................................... ..................................... 30
23. M methods of placing top dress fertilizer .......... ....................................... ................ 31
24. M anagem ent of crop residues .................................................... ................................ ..... 32
25. Common maize pests reported by sample farmers.......................................... ............... 32
26. Transportation of maize from field to homestead ...................................... .................. 33
27. Methods of grain storage across zones .................. ......................................... 33
28. Seed selection criteria .................... ................. .... .... .... ........ ........... ..... 34
29. Mean number of years of recycling improved cultivars ................................................... 34
30. Credit availability, credit sources, and problems of getting loans ............................................... 37
31. Sources of maize production t. 1I 111d ,_. information for the lowland zone .............. .................... 38
32. Sources of maize production t. 111 1 .- information for the intermediate zone ............................ 38
la. Rainfall pattern for the highland zone, monthly totals. ........................................ ............... 3
lb. Rainfall pattern for the intermediate zone, monthly totals. ........................... ............................. 3
lc. Rainfall pattern for the lowland zone, monthly totals. ...................................... ............... 3
Id. M ean annual rainfall for selected stations in each zone. .............................. ............................... 3
2. Northern Zone villages sampled for survey, 1995. ......................... ........ .......................... 4
3. Farm size and maize area in the lowland zone. ............................... .... ...................... .... 13
4. Farm size and maize area in the intermediate zone. ...................................................... 13
5. Maize cropping calendar for agroecological zones in northern Tanzania ..................................... 17
6. Adoption of maize varieties by agroecological zone, 197494. ............... ................ ........... 19
7. Adoption of maize varieties by land preparation method. ...................................................... 19
8. Local and improved maize yields by land preparation method. ................................................ 19
9. Local and improved maize yields by zone. .......................................................... .............. 19
10. Fertilizer application by land preparation method. .............. ...... ...................... ............. 29
11. Fertilizer application by agroecological zone................... ................................. 29
ABBREVIATIONS AND ACRONYMS
CAN Calcium Ammonium Nitrate
CIMMYT Centro Internacional de Mejoramiento de Maiz y Trigo [International Maize and Wheat
DRT Department of Research and Training
FSD Food Security Department
FSR Farming systems research
ICW Ilonga Composite White
IMR Inverse Mills Ratio
masl Meters above sea level
MDB Marketing Development Bureau
MOA Ministry of Agriculture
MSV Maize streak virus
NGO Non-governmental organization
NMRP National Maize Research Programme
n Sample size
OLS Ordinary Least Squares
OPV Open pollinated variety
Ox (or O) Oxen
SA Sulphate of Ammonia
SACCAR Southern Africa Center for Coordination of Agricultural Research
SARI Selian Agricultural Research Institute
SG2000 Sasakawa-Global 2000
ST Streak resistant
SUR Seemingly unrelated regression
TARO Tanzania Agricultural Research Organization
TMV Tanzania maize variety
Trac (or T) Tractor
Tsh Tanzanian Shillings
UCA Ukiriguru composite A
ULVA Ultra low volume applicators
USAID United States Agency for International Development
WAP Weeks after planting
The Maize Research Impact Evaluation study is the culmination of the efforts of many people and
institutions. This makes the task of recognizing each contributor exeedingly difficult. We are especially
grateful to the northern Tanzania maize farmers who participated as our sample subjects. Despite the
difficult charge of simultaneously managing their farms and raising families, they took the time to respond
to our rather lengthy questionnaire. Their time and candid responses made this study possible. We are
indebted to SACCAR, CIMMYT, and the Government of Tanzania for financial support. We are grateful to
Dr. Joel Ransom of CIMMYT for financial and logistical support for the survey.
Special thanks go to CIMMYT, which for the past 20 years has collaborated with the Tanzanian National
Maize Research Program. CIMMYT has made a tremendous contribution to the advancement of maize
production technologies in the country.
We wish to recognize the contribution of Selian FSR staff, especially Mr. Peter Sulumo, Mrs. Owenya, and
Mrs. Modestus, who helped interview farmers and compile and clean data. We also thank Mr. Mariki of
SARI and Mr. Njau, an extension officer from Arusha region who participated in the survey. Our special
acknowledgment goes to Dr. Moshi, the Zonal Research and Training Director (Eastern Zone) for
coordinating the national maize research impact study. His tireless efforts and commitment made this study
possible. We also wish to thank Dr. Haki for his able leadership and for providing the human resources and
logistical support required for the survey and data analysis.
Maize provides 60% of the dietary calories and more than 50% of utilizable protein to the Tanzanian population.
The crop is cultivated on an average of two million hectares, which is about 45% of the cultivated area in
Tanzania. Realizing the importance of the maize crop to the lives of Tanzanians, the government has been
committing human and financial resources to develop the industry. A National Maize Research Program (NMRP)
was started in 1974 with the broad objective of developing cultivars suitable for major maize producing areas. The
NMRP and maize extension services have made considerable impact on increasing food production.
The objective of this study was to evaluate the impact of maize research and extension for the past 20 years. It
was conducted by the Department of Research and Training (DRT) in collaboration with the Southern Africa
Coordination Center for Agricultural Research (SACCAR) and the International Maize and Wheat Improvement
Center (CIMMYT). To increase data validity and reliability, farmers were interviewed by researchers and
experienced extension officers using a structured questionnaire. Interviews were conducted in all seven
agroecological zones of the country between June and November 1995. This report covers survey findings in the
Northern Zone, which includes Arusha and Kilimanjaro regions. Northern Tanzania is an important maize
growing area, accounting for about 10% of the total national maize production. The zone is one of the country's
maize surplus areas. Total area under maize production in the zone is 160,700 ha, of which 70% is in the Arusha
Data collected in the survey were grouped into two major agroecological zones: the intermediate and lowland sub
zones, the most important zones for maize production in the study area. Analysis was also undertaken after
sample farmers were categorized according to their method of land preparation. A two-sample comparison
between zones and among methods of land preparation was employed to investigate statistical differences. A two
step Heckman's procedure was used to simultaneously analyze factors affecting adoption of improved maize seed
and inorganic fertilizer.
Results of the analysis showed that raising livestock in the zone was quite common. Of the 126 farmers sampled,
84% kept cattle, 79% kept goats, and only 40% raised sheep. Overall, each household averaged about five head
of cattle, six goats, and two sheep. All sampled households had hoes equivalent to the number of family laborers
available. Ox-ploughs were for land preparation and planting in Arumeru, Babati, Hanang, and Mbulu districts in
the Arusha region and Hai District in the Kilimanjaro region. The decision to preparare land by hand-hoe was
largely dictated by terrain. Rolling topography in the highlands makes mechanization difficult. In the intermediate
zone, tractors were the major means of land preparation. The majority of farmers owning tractors were found in
Intercropping maize with beans or pigeon peas is the most common cropping system in northern Tanzania. Sixty
percent of sample farmers in the intermediate zone grew maize in association with beans and pigeon peas. Forty
percent of sample farmers in the lowland zone grew maize in pure stand. Coffee in association with banana is a
common cropping system on the mountain slopes where rainfall is high and temperatures are cool. About 19% of
the sample farmers grew coffee in association with banana.
All sample farmers reported that they used either certified or recycled improved seed. On average, about 80% of
the sample farmers recycled improved maize seed for 4 to 6 years. Recycling improved maize seed, including
hybrids, has reduced the adoption of improved maize seed in the two zones. Extension agents should discourage
farmers from recycling hybrids and advise them on how to recycle composites. CG4141 is the most preferred
hybrid in both zones because of its yield and drought tolerance/resistance/avoidance attributes. Forty-five percent
of sample farmers at one point had discontinued growing an improved variety for various reasons. The varieties
that have been discontinued by most farmers in the intermediate zone are hybrids H622 and H632, and Kilima
and Katumani in the lowlands. The major reason given for discontinuing hybrids was their late maturity.
Over the span of the study, there was no statistically significant difference (P=0.05) of rate of adoption of
improved maize varieties among methods of land preparation except for hand-hoe versus oxen in 1990, and
hand-hoe versus tractor in 1990 and 1991. Comparing the intermediate and lowland sub-zones of the Northern
Zone showed that only farming experience significantly affected (P=0.05) the probability of adopting improved
maize seed in the zone. An increase in one year of experience increased the probability of adopting by 1.3%. The
reason other factors did not have a significant impact on probability of adoption could be the high rate of
adoption of improved maize seed. Also no factor significantly influenced (P=0.05) the intensity of adoption of
improved maize seed. The reason for the non-significant difference may be the narrow range of intensity of
adoption (0.17 1.0 ha, with a mean of 0.89 and a standard deviation of 0.21 ha). Farmers reported that the
major non-household factors affecting adoption of improved maize varieties were prices of inputs, poor marketing
systems, and varietal traits.
The adoption of agronomic technologies was generally high. Farmers reported that crop planting time is dictated
by the onset of rainfall. Hence, the significant difference in timing of planting (P 0.01) is a factor of location (and
onset of rainfall), rather than management differences. Row planting was practiced by all but one of the 126
respondents. The average plant population was slightly lower than the recommended population of 44,444
plants/ha for the medium and longer duration varieties commonly grown in the zone.
About 89% and 60% of sample farmers in the intermediate and lowland zones, respectively, weeded their maize
twice during the growing season. The difference in the proportion of farmers weeding twice between the two
zones is significant at p = 0.01. It may be explained by the length of the growing season. In the lowlands, the
growing season is shorter so farmers may not have time to weed twice. Only 21% of the sample farmers in the
lowland weeded their maize fields once and followed the recommended time for the first weeding. Farmers in the
intermediate zone did their first weeding 3.8 weeks after planting, compared to 3.6 weeks for the lowland
farmers. About 80% and 92% of sample farmers in the intermediate and lowland zones, respectively, applied
fertilizer. The difference in the percentage of farmers applying fertilizer between the two zones is significant at p =
0.01. Application of organic manure was more common in the Arusha region than in Kilimanjaro. About 49%
and 44% of sample farmers in the intermediate and lowland zones, respectively, applied kraal manure. While
64% of the farmers in the intermediate zone applied chemical fertilizer, only 44% of sample farmers in the
lowland zone used chemical fertilizer in the 1994/95 season. The difference in the proportion of farmers
applying chemical fertilizer is significant at P=0.05. Up until 1994, the fertilizer application level was still less than
half the recommendation for intermediate and highland areas, which averaged 80 kg N/ha. The majority of
farmers reported that they top dressed fertilizer only once. Only 10% of sample farmers applied a second
application of top dressing fertilizer, while only 4% applied basal chemical fertilizer at planting. On average, timing
of fertilizer application was close to the recommendation. Applications ranged from 1 to 13 weeks after planting
maize, with an average of 4.7 weeks. There was no significant difference (at p = 0.05) between zones on time of
fertilizer application. The majority of farmers in the intermediate zone broadcast fertilizer, contrary to the
recommendation of spot or furrow application and covering fertilizer with soil. In the lowland zone, the majority
of farmers banded fertilizer around crops without covering it. The recommendation for fertilizer placement was
followed poorly in both zones.
For both zones, units of livestock significantly influenced the probability of adopting fertilizer (p = 0.05). A unit
increase in livestock units decreased the probability of adopting fertilizer by 1.5%. The reason may be that
farmers with large herds have more manure, which they can substitute for chemical fertilizer. In addition, farmers
who have large cattle herds live in areas with less rainfall, hence less soil leaching, which makes their soils more
fertile and lessens the need to apply fertilizer.
For the two zones, no variable significantly influenced (p=0.05) the intensity of adoption of fertilizer. However, for
the intermediate zone, livestock herd size significantly influenced the intensity of fertilizer adoption (p 0.01). The
income effect of livestock herd size may have contributed to its positive influence, as opposed to its negative
impact on probability to adopt fertilizer. Richer farmers, as manifested by larger livestock herds, were likely to use
higher doses of fertilizer. The crop residue management recommendation is to plough it under in order to avoid
soil mining for farmers who do not apply or apply only small amounts of fertilizer. However, only 11% of the
farmers followed this recommendation. About 94% of the respondents from the lowland zone, compared to 71%
from the intermediate zone, reported that they fed their maize stover to cattle in the field. The majority of farmers
in the lowland zone do not zero graze because they have large herds of cattle. Only 14% of the farmers cut and
carried stover home for feeding zero-grazed dairy cows. This results in nutrient flow from the field to other areas.
Extension officers should encourage farmers who cut and carry stover and those who feed crop residue in situ, to
apply fertilizer to their fields in order to replenish the exported nutrients.
Stalk borer (Buseola spp.) is one of the most important maize pests in Tanzania. As expected, a significantly
higher percentage of farmers in the lowland zone have been affected by stalk borers as compared to farmers in
the intermediate zone. About 58% and 20% of the farmers in the intermediate and lowland zones, respectively,
used chemical control against insect pests. Due to efforts aimed at breeding for disease control, maize diseases
were rarely reported by sample farmers. The major disease reported was Maize Streak Virus (MSV). No farmer
reported using any measure to control the disease.
Most farmers in Kilimanjaro and the lowlands of the Arusha region store their maize using gunny bags or airtight
drums. For sample farmers who used gunny bags, 80% treated maize using chemicals alone or in combination
with ashes. Farmers who used drums did not use chemicals and had no grain losses. Treatment with Actellic
Super powder for shelled maize was the most common chemical control. Upright cribs are used in Arusha.
Farmers who used cribs normally did not shell their maize. Ninety percent of tractor users treated their shelled
maize as compared to 77% of hand-hoe users.
Sixteen percent of the sample farmers in the two zones received a loan. Sasakawa-Global 2000 (SG-2000) was
the major source of credit for small-scale, subsistence farmers in northern Tanzania. The major disadvantage of
the SG-2000 credit service was its short-term nature (3 years) and weak credit administration that increased the
loan default rate when it was extended to a large number of farmers. About 19% of sample farmers in the
intermediate zone received credit once as compared to 8% of the farmers in the lowland zone. The difference in
loan accessibility may be due to the fact that SG-2000 had more sites in the intermediate than in the lowland
zone. Only 1 out of 3 loan recipients in the lowland zone, compared to 1 out of 13 in the intermediate zone,
reported that they received a loan from a formal bank. Lack of collateral and cumbersome loan application
procedures were the major obstacles that farmers from both zones faced in securing loans from formal credit
institutions. The three most important sources of maize production information were extension agents, farmers,
and non-governmental organizations. In both zones, extension officers were the most important sources of all
agricultural production technologies, except for technologies pertaining to draft animals in the lowlands.
1.1 Introduction to the study
Maize is the major cereal consumed in Tanzania. It is estimated that the annual per capital
consumption of maize in Tanzania is 112.5 kg; national maize consumption is estimated to be three
million tons per year. In the Northern Zone, the per capital consumption is estimated to be 130 kg
per year. Maize contributes 60% of dietary calories to Tanzanian consumers (FSD 1992, 1996). The
cereal also contributes more than 50% of utilizable protein, while beans contribute 38% (Due 1986).
Maize is grown in all 20 regions of Tanzania. The crop is cultivated on an average of two million
hectares or about 45% of the cultivated area in Tanzania. However, most of the maize is produced in
the Southern Highlands (46%), the Lake Zone, and the Northern Zone. Dar es Salaam, Lindi,
Singida, Coast, and Kigoma are deficit regions. Dodoma is a surplus region in good growing years,
and the region is the number one supplier of maize to Dar es Salaam in years following a plentiful
rainfall season (FSD 1992; Mdadila 1995).
Maize is not only a staple crop in surplus regions, it is also a cash crop. For instance, in the Lake
Zone, maize competes aggressively with cotton for land, labor, and farmers' cash. Realizing the
importance of the maize crop to lives of Tanzanians, the government has been committing human
and financial resources to develop the industry. Research and extension efforts in maize started in
1960. Breeding efforts in the 1960s resulted in the release of Ukiriguru composite A (UCA), and
Ilonga composite White (ICW). Between 1973 and 1975 Tanzania experienced a severe food
shortage due to drought and the "villagization" campaign that displaced farmers (Maliyamkono and
Bagachwa 1990). The food crisis prompted the nation to launch several campaigns such as
agriculture for survival" (kilimo cha kufa na kupona) with the objective of food self-sufficiency. The
country also launched a maize project in 1974 with assistance from the U.S. Agency for International
Development (USAID). Its objective was to promote maize production in pursuit of food self-
sufficiency. On the research frontier, the National Maize Research Program (NMRP) was launched
with the broad objective of developing cultivars suitable for major maize producing areas.
The NMRP and maize extension have made considerable impact on increasing food production. This
study was conducted to evaluate the impact of maize research and extension during the past 20 years.
Conducted by the Department of Research and Training (DRT) in collaboration with the Southern
Africa Coordination Center for Agricultural Research (SACCAR) and the International Maize and
Wheat Improvement Center (CIMMYT), the study included the nation's seven agroecological zones.
The study was conducted between June and November 1995. This report covers the survey findings
in the Northern Zone.
1.2 Objectives of the study
The objectives of the study were to:
describe the maize farming systems in northern Tanzania,
evaluate adoption of maize production technologies in the Northern Zone, and
define the future research agenda in light of the study's findings.
1.3 Description of the study area
Northern Tanzania consists two of mainland Tanzania's twenty regions-Arusha and Kilimanjaro.
The Arusha region alone covers 82,000 km2, about 9.3% of the area of Tanzania. With a population
density of 17 persons/km2, Arusha is one of the country's most sparsely populated regions (MDB
1993). It has a diverse climate that enables it to produce a wide range of crops and it is the number
one wheat producer in the country, accounting for more than 64% of national wheat production
each year since 1981. The region was the second largest producer of maize and beans between
1981 and 1989 (Nkonya et al. 1991). Arusha is also endowed with the world's richest game parks
that attract tourists from all over the world.
The Kilimanjaro region has an area of 13,000 km2, which is 1.5% of mainland Tanzania. With a
population density of 94 persons/km2, Kilimanjaro is the third most densely populated region in the
country after Dar es Salaam (1,849 persons/km2) and Mwanza (97 persons/km2) (MDB 1993).
Kilimanjaro is a leading producer of arabica coffee in the country, producing about 40% of its mild
arabica coffee (Nkonya et al. 1988). The Kilimanjaro region also attracts tourists drawn by the
highest mountain in Africa, snow-capped Mt. Kilimanjaro, that stands at 19,840 ft (5,890 m) above
The northern Tanzania climate also allows production of temperate and tropical high value crops,
specifically, flowers, snap beans, barley, garlic, pigeon peas, paddy, and onions. Excluding Dar es
Salaam, northern Tanzania has the best communication system in the country. There is an
international airport, about 700 km of paved road, and a railroad that connects the zone with two
important ports, Tanga and Dar es Salaam.
Northern Tanzania is an important maize growing area that accounts for 10% of the total national
production of the cereal (Nkonya et al. 1991) and is one of the nation's maize surplus areas. Total
area under maize production in the zone was 160,700 ha, of which 70% was in the Arusha region
(MDB 1993). The major maize producing districts are Mbulu, Babati, Hanang, and Arumeru; other
maize producing districts of less importance are Moshi, and Rombo. Remaining districts-Mwanga,
Same, Kiteto, Monduli, Ngorongoro, and Simanjiro-are maize deficit areas because of their
The Northern Zone has three major agroecological zones:
(1) High Rainfall Zone: This zone receives about 1,200 1,500 mm of rainfall per year. Rainfall
distribution is good and reliable. The zone is located on the slopes of Mt. Kilimanjaro, Meru,
Hanang, Monduli, Pare, and Ngorongoro mountain ranges; Oldeani and Loolmalasin. Some
areas located in the high plateaus fall into the high rainfall zone including Bashnet in Babati and
Mama Isara area in Mbulu. Most areas in the high rainfall zone rise to an altitude of 1,500
meters above sea level (masl). Other areas are above 1,500 masl, but their rainfall is less than
1,200 mm per annum (these areas are always on the lee side of mountains, e.g. Olkokola in
Arumeru). Major crops grown in this high rainfall zone are coffee in association with banana.
(3) Low Rainfall Zone: This zone receives rainfall ranging from 500 to 800 mm per year with
very erratic distribution. Low rainfall areas are always in the lowland plains below 900 masl.
Studies show that drought occurs in one out of every four years. Because there is no land
pressure in this zone, farmers in the high and moderate rainfall zones grow their maize and
other annual crops in this area, the second most important area for maize production in
northern Tanzania. The major cropping systems in the zone are monocropped maize,
monocropped beans, and maize in association with beans. Extensive livestock-keeping prevails
in the zone, which is the most important area for livestock production in northern Tanzania.
Land preparation is predominantly accomplished using ox-plough and tractor (Cunard et al.
1983; Nkonya et al. 1991).
1.4.1 Sampling procedure
This report is part of a national survey covering all agroecological zones of Tanzania. The number of
farmers interviewed in each zone was determined by the importance of maize production in the area.
About 1,000 maize farmers were interviewed nationwide. The Northern Zone was allocated 126
farmers, or approximately 13% of the national sample. At the zonal level, farmers were sampled
from districts with significant maize production based on figures from the statistical unit of the MOA.
Seven districts were purposively sampled. At the district level, villages were selected with the help of
extension staff. Seven villages were purposively sampled (Figure 2). From each village, approximately
18 farmers were randomly sampled from the register of households. To increase data validity and
reliability, farmers were interviewed by researchers and experienced extension officers using a
structured questionnaire developed by a panel of the zonal farming systems research economists,
CIMMYT, SACCAR economists, and national maize breeders and agronomists. The interviews were
conducted between June and November 1995.
To maintain uniformity, data from all zones were
compiled at Selian Agricultural Research Institute
(SARI) and then returned to the respective zones
for analysis and completion of the zonal reports. awate
MIangarini *Kin i
1.4.2 Data analysis Shimbi
Data were grouped into two major 4\
agroecological zones, the intermediate and
lowland zones, which were the most important Managa
zones for maize production in the study area.
Even highland zone farmers grew maize in either Gocho
the intermediate or lowland zones. Sample
farmers were also grouped according to the
method of land preparation they used. Two-
sample comparisons between zones and among
methods of land preparation were made to
investigate statistical differences. A two-step Figure 2. Northern Zone villages sampled for survey, 1995.
Heckman's procedure was used to analyze factors affecting the probability of adopting improved
maize seed and inorganic fertilizer.
This procedure was used because it addressed simultaneity problems. Farmers tend to adopt
improved maize seed and fertilizer simultaneously: farmers who plant improved maize seed are likely
to apply fertilizer. The converse is also true: farmers who apply fertilizer are likely to plant improved
maize seed (Smale et al. 1994; Nkonya et al. 1997; Kaliba et al. 1998). Hence, when estimating the
probability to adopt maize seed, inorganic fertilizer should be included as an independent variable in
the model. Also, when estimating the probability of using inorganic fertilizer, improved maize seed
should be included as an independent variable. When simultaneously determined variables are
included as independent variables in a system of (simultaneous) equations, they are referred to as
endogenouss" variables. The adoption of improved maize seed and fertilizer is estimated in a system
of two simultaneous equations, one for each endogenous variable. The Heckman's two-step
procedure involves two estimation steps. In step 1, a probit equation is estimated for each of the
simultaneous equations. The probit model is used because its likelihood function is well behaved as it
gives consistent Maximum Likelihood Estimate (MLE) coefficients (0) and the standard error of the
estimate (o) (Maddala 1983). The estimated probit models are:
P(AS)= f(FARMS,EXP,EDUC,LUNIT,LABOR,MLP1,MLP2,MLP3,NRATE,Ui) .......................... (1)
P(AF) = f(FARMS,EXP,EDUC,LUNIT,LABOR,MLP1,MLP2,MLP3,ADIS,Ui)............................. (2)
P = probability;
AS = 1 if farmer uses improved maize seed, 0 otherwise;
AF = 1 if farmer uses fertilizer, 0 otherwise;
FARMS = Farm size (acres);
EXP = Farming experience (years);
EDUC = Level of education of family head (years);
LUNIT = Livestock units (index where livestock numbers are aggregated using following weighting
factors: cow = 0.8; goat = 0.4; sheep = 0.4);
LABOR = Family labor (index where family members are aggregated using following weighting
factors: male and female adults above 16 years = 1; children 12-15 years = 0.5);
MLP 1 = Hand-hoe (number);
MLP2 = Ox-plough (number);
MLP3 = Tractor (number);
MLP 4 = Other method (number);
NRATE = Rate of nitrogen (N) applied (kg/ha) endogenouss variable);
ADIS = Number of acres planted with improved maize seed endogenouss variable); and
Ui = Random error
Results of the first step show the influence of independent variables on the probability of adopting a
given technology (ap/ax). The results are the same as any other probit model estimated as a single
In step 2, the intensity of adopting fertilizer or improved maize seed is estimated as follows:
E(yi yi>0) = 3x + of(3xi)/F(p xi) .............................................. ..................... ... (3)
where yi = intensity of adoption of a technology (area under improved maize seed or level of
fertilizer), ith = independent variable as specified in equations 4 and 5 below.
The ratio of f(3px)/F(3px) is the Inverse Mills Ratio (IMR), evaluated at each sample observation. IMR is
calculated from the probit results of the first step. Using the data from adopters only, an Ordinary
Least Squares (OLS) model, including the IMR as regressor, is estimated for each endogenous
variable. The impacts of the same factors on intensity of adoption (aADIS/aX and aNRATE/aX) are
estimated using a system of Seemingly Unrelated Regression (SUR) equations. In our case, the
following two OLS equations will be estimated simultaneously in a system of equations:
ADIS = f(FARMS,EXP,EDUC,LUNIT,LABOR,MLP 1,MLP2,MLP3, NRATE IMRS,Ui)....... (4)
NRATE = f(FARMS,EXP,EDUC,LUNIT,LABOR,MLP1,MLP2,MLP3,ADIS,IMRF,U) ............ (5)
The variables are as defined earlier.
There are two advantages to using the Heckman's approach. The first is that we use SUR1 in the
second stage because only adopters are included in the second stage analysis. SUR is easy and
familiar, hence available in many econometric computer packages. The second advantage is that the
use of SUR purges the heteroscedasticity of the estimates obtained from the second step (Greene
1993). Some of the factors influencing adoption were either not variables across farmers or could
not be collected from household surveys. Those variables were analyzed using cross tabulations.
A SUR is a system of equations whose individual equation random errors are believed to have some correlation.
For instance, in our case, errors of equation estimative intensity of adoption of maize seed are believed to be
correlated with errors of the equation estimating intensity of adoption of fertilizer. In this situation, the equations
are estimated as a system and not individually, in order to account for the cross equation relationship of random
2.0 HISTORY AND DEVELOPMENT OF MAIZE RESEARCH
About 85% of Tanzania's total maize production is grown by peasants whose farms are less than 10
ha. Ten percent of maize is produced on medium-scale commercial farms (10-100 ha), and the
remaining 5% is grown on large-scale commercial farms (over 100 ha). Between 1961-65 and
1985-95 the growth rate of maize production in Tanzania was estimated to be 4.6%, of which 2.4%
and 2.2% per year were due to growth in area and yield, respectively. Despite this growth in yield,
the national average is less than 1.5 t/ha, however, grain yields are higher in high potential areas
such as the southern highlands (Moshi and Marandu 1988).
Maize breeding and agronomy trials have been conducted in Tanzania for more than 20 years. The
improved open pollinated varieties (OPVs), namely ICW and UCA, were developed, tested, and
released in the 1960s and are still widely used. During the same period a few research stations
undertook agronomic research, which later formed the basis for recommendations that were applied
to the entire country.
In 1974, the NMRP was launched as a means to coordinate maize research, which included better
utilization of some resources. The program is responsible for coordinating all phases of maize
research, from varietal development and maize management research on station, to verification in
farmers' fields. The program divided the country into three major varietal recommendation
agroecological zones: (i) highland zone (elevations above 1,500 masl), with a growing period of 6-8
months; (ii) intermediate zone (900-1500 masl), which is further subdivided into 'wet' (>1,100 mm
rainfall) with a 4-5 month growing period, and 'dry' (<1,100 mm rainfall) with a 3-4 month growing
period; and (iii) lowland zone (0-900 masl) with a 3-4 month growing period. To date several
breeding populations have been developed and are being improved through recurrent selection for
specific traits. Since 1974, two hybrids and six OPVs have been released. In 1976, Tuxpefo was
released for the lowland areas. Hybrids H6302 and H614, suitable for the highland zone, were
released in 1977 and 1978, respectively. In November 1983, three OPVs were released: Kito,
Kilima, and Staha. Staha is characterized by its tolerance to maize streak virus disease, whereas
Kilima was recommended for the mid-altitude zone. Kito is an early maturing variety adapted to both
low and mid-altitude zones. In 1987 two OPVs, TMV1 and TMV2, were released. TMV1 is white
flint, streak resistant, and has intermediate maturity. It is recommended for the lowland and mid-
altitude zones. TMV2 is also white flint and is recommended for the high altitude, high-potential
maize producing areas.
In 1994, the NMRP released versions of Kilima, UCA, Kito, and Katumani that are resistant to maize
streak virus: Kilima-ST, UCA-ST, Kito-ST, and Katumani-ST, respectively. Around the same time two
foreign seed companies, Cargill and Pannar, introduced/released seven hybrids for commercial use
by the farmers. For the improvement of husbandry practices, the NMRP conducted off-station
agronomic trials that resulted in the 1980 recommendations for maize production practices specific
to 11 regions. The recommendations were on varieties, spacing, plant density, fertilizer rate,
weeding regime, and pesticide use. Maize research work in the Northern Zone was formerly
coordinated from the Lyamungu Agricultural Research Institute. Research activities focused on soil
fertility, weed control, cropping systems, and variety testing. Until early 1994 these research
activities were collaboratively conducted by staff from the maize section, the FSR program, and the
National Bean Research Program.
In 1994 the coordination of maize research activities in the Northern Zone was moved to SARI,
which is also the zonal headquarters for research and training activities. Aside from the activities they
formerly coordinated at Lyamungu, the maize staff was also made responsible for:
1) Initiating maize breeding activities to cater to the mid-altitude areas of the country.
2) Collaborating with staff at Arusha Foundation Seed Farm at Ngaramtoni to maintain purity of
existing maize inbred lines.
3) Initiating activities on inbred lines from the existing OPVs.
4) Performing crosses among existing inbred lines as well as among inbred lines to be extracted
from existing OPVs, in order to obtain new maize hybrids for mid-altitude areas
(Anandajayasekeram et al. 1992).
2.2 Maize production technology recommendations
As indicated, most maize growing areas in Tanzania are divided into three agroecological zones: the
low altitude zone (less than 900 masl); the medium altitude zone (900-1,500 masl); and the high
altitude zone (above 1,500 masl). Maize production recommendations were developed to fit the
agroecological zones as outlined below.
The choice of maize variety is determined by the farmer's objective, the length of growing season,
the elevation, and the amount of rainfall. Recommended varieties for the various ecological zones in
Tanzania, including those of the Northern Zone, are shown in Table 1.
2.2.2 Planting time, method, and spacing
Generally, early sowing is the most important single factor for increased grain yield (Goodbody
1990). Lower yields and increased pest and disease attack occur with late sowing. Varieties
susceptible to maize streak virus disease suffer more when sown late.
The total length of the growing season and the number of days to silking of varieties influence the
decision on when and what variety to grow. If the rains come late, farmers are advised to plant early
maturing varieties even in areas that are suitable for full season varieties.
In most areas of the Northern Zone maize is normally sown between mid-January and mid-March. In
isolated pockets that have bimodal rainfall,-Rombo, Mwanga, and Same districts in Kilimanjaro
region-farmers also sow their maize in October.
It is recommended that seed be sown directly into moist soil to protect it against rodents, birds, and
drying. Normally 5-7 cm is considered an adequate sowing depth. Deeper sowing retards
emergence. In dry areas, maize seed may be sown deeper and then covered with soil.
The best way to get uniform stands is to sow in regularly spaced rows and at regular intervals within
the row. Row planting makes weeding and insect control easier. Generally, across all of the Northern
Zone's agroecological zones plant populations range from 40,000 to 53,000 plants/ha. For small
statured varieties, namely, Kito and Katumani, optimum plant density can range from 60,000 to
70,000 plants/ha (Matowo and Mgema 1990b). Taking into account unreliable seed quality, insect
pests, and vermin it is advisable to sow more plants per hill. However, excess plants (if any) need to
be thinned about 2-3 weeks after seedlings emerge from the soil.
Recommended spacings for full season varieties (H6302, H614, H622, Kilima/Kilima ST, UCA/
UCA ST, ICW, Tuxpefo, Staha, Pannar and Cargill hybrids) are 75x30 cm or 90x25 cm with a final
stand of one plant per hill. Using this spacing, a plant population of 44,444/ha is achieved. More
recent results show that in areas above 1,500 masl with reliable rainfall, two plants per hill at 90x50
cm and three plants per hill at 90x75 cm gave similar results to the spacing cited above. In the drier
intermediate altitude areas and the lowland and coastal zones, two plants per hill at 75x60 cm gave
the same yield as one plant per hill at 75x30 cm. For small statured varieties (Kito and Katumani) a
spacing of 75x40 cm with two plants per hill is recommended.
Table 1. Commercial varieties in Tanzania and their potential
Lowland and Coastal areas (0-900 m)
Ilonga Composite (ICW)
Low rainfall (below 1,000mm)
High rainfall (over 1,000mm)
High altitude areas (over 1,500m)
Source: Anon. (1988).
Expected yield under
good husbandry (t/ha)
2.2.3 Fertilizer type, time, and method
To provide nitrogen one can use either urea, calcium ammonium nitrate (CAN), or sulphate of
ammonia (SA). Nitrogen application may be split, with 30-50% of the total requirement applied at
planting. The remaining N should be applied when maize is about a meter high. Phosphorus (P205)
is necessary to promote root growth, strong stems, and good grain. The entire recommended
amount should be applied at sowing. Table 2 summarizes the fertilizer recommendation for northern
For lowland plains, 40kg N/ha is recommended. For areas receiving rainfall above 800 mm per
annum, 80 to 112 kg N/ha is recommended. No response for triple super phosphate (TSP) has been
observed, but for high potential areas a basal application of 0-40 kg P205 (for replenishing
phosphates used up by crops) is recommended.
Fertilizer is normally placed 5 cm below the depth of the seed and about 5 cm to the side. This is
accomplished by digging a single hole beside each seed and placing fertilizer in the hole and covering
it with soil. Alternatively, a continuous furrow can be made along the length of the planting row.
Fertilizer is then placed in the furrow and covered with soil. The seed is then planted on top of this
soil and covered properly.
2.2.4 Weeding time, frequency, and method
Weed control in maize is important to reduce competition for water, soil nutrients, and light.
Research results have consistently shown that late and poor weeding can result in yield reductions of
30% to 70% (Matowo and Mgema 1990a). Generally, two hand-weedings should suffice at all
elevations. However, durations between the first and second weeding vary between locations. For
example, for areas over 900 masl weeding should be done at 2 and 5 weeks after planting (WAP).
For areas under 900 masl weeding should be done at 2 WAP and 4 WAP. Weeds may also be
controlled by using various herbicides. The following herbicides have been recommended for use in
1) Atrazine (Gesaprim)
2) Atrazine + metalachlor (Primagram)
3) Alachlor (Lasso/atrazine)
4) Pendimethalin (Stomp)
Table 2. Fertilizer recommendation for maize according to agroecological zones
Fertilizer rate (kg/ha)
Altitude (masl) Rainfall N P20O Districts covered
0-900 High 20-45 20 Mos Hai Arm Rom
Low 0-20 0-20
900-1500 High 40-112 20-40 Han Mbu Mos Hai Rom Mond Arm
>1500 High 40-112 20-40 Mos Hai Rom Arm
Low 20-50 20
a Mos = Moshi, Arm = Arumeru, Mbu = Mbulu, Han = Hanang, Rom = Rombo, and Mond = Monduli.
Source: Haule (1988); Samki and Harrop (1984); Mowo et al. (1993).
On the other hand, in maize/bean intercropping alachlor plus linuron (Lasso/linuron) and
metabromuron + metoalachlor (Galex) have been recommended.
2.2.5 Pest and disease control
Stalk borers and armyworm are the two major insect pests of maize. Stalk borers can be controlled
fairly easily with endosulfan, malathion, carbaryl, and sumithion when applied at the correct time.
When the plants have about seven leaves, a small amount of dust should be applied into the funnel of
the leaves. About two weeks after the first dusting, a second application should be made. Armyworm
outbreaks in Tanzania can be effectively controlled only with intervention by national or international
organizations, as aerial insecticide applications are required in outbreak areas. Individual farmers,
however, can respond to lower-scale infestations with knapsack sprayers and ultra low volume
applicators (ULVA). Malathion, fenitrothion, permethrin, and endosulphan are all very effective.
There are five major maize leaf diseases: common rust caused by Puccinia sorghi; lowland rust
caused by Puccinia polysora; Helminthosporium turcicum; Helminthosporium maydis; and maize
streak virus disease (MSV). None of these diseases can be economically controlled by chemical
means. Biological control through breeding for disease resistance or tolerance is the only economical
There are also three common cob diseases: Gibberella spp., Fusarium spp., and Diplodia spp.
These diseases attack the grain and the cob. To break the disease cycle, all diseased cobs should be
destroyed at harvest and diseased plants and insects should be burned.
3.0 SOCIOECONOMIC AND DEMOGRAPHIC CHARACTERISTICS
3.1 Demographic characteristics
Table 3 summarizes the family characteristics of households in northern Tanzania. Most respondents
were middle aged (in their early forties). Farmers in the lowland zone tended to be younger than the
intermediate zone respondents, however, the age difference was not significant (p = 0.05). Early
crop farmers in northern Tanzania settled on the fertile slopes of Mt. Kilimanjaro, Meru, Hanang,
and other mountains. Land pressure pushed younger farmers to settle in the lowlands where rainfall
is unreliable (Dunford 1980). Their farming experience is about 19 years, implying that they started
farming at 20 years of age. In both zones, the mean for family labor is four, 50% of which is female
labor. Family labor may not be enough to accomplish all operations during peak periods of planting,
weeding, and coffee picking.
About 56% of sample households hired labor for weeding and harvesting maize and beans and
picking coffee. Among farmers who hired labor, 86% hired labour for weeding maize and 23% for
harvesting. Eighty percent of sample farmers who grew coffee hired labor for harvesting the crop.
The level of education of the household heads was 6.5 years in the lowland zone and 5.8 years in
the intermediate zone. As expected the younger farmers in the lowland zone were slightly better
educated than their peers in the intermediate zone. No sample farmer was illiterate. A number of
farmers did not have formal education, but had attended adult education classes. In such cases, a
maximum of four years of formal education was assigned to farmers to reflect their literacy level.
3.2 Land resources and allocation patterns
Land is a limiting factor in northern Tanzania. This is especially true on the Kilimanjaro and Meru
mountain slopes. The land shortage results in small plots and a large number of holdings. In the
plains of almost all districts of the Arusha region, land shortage is not as acute. Since Arusha is
sparsely populated, farmers in the region generally have bigger farms than in the Kilimanjaro region.
Survey findings confirm this observation as farmers sampled from Arusha had bigger land areas than
those from Kilimanjaro. Figures 3 and 4 show an upward trend of total farm size and maize area in
both zones. The reason for this trend may be that most farmers started farming 20 years ago (the
mean farming experience of sample farmers is 19.4 years with a standard deviation of 11 years).
When farmers start farming, most have small
holdings. They normally increase their farm size Table 3. Demographic characteristics of sample
through buying, renting, and clearing land in households
areas with surplus land like Hanang, Babati, and Lowland Intermediate
some Mbulu District areas. However, maize area Mean of: (n=36) (n=90) P-valuea
in the lowlands dropped in 1984. Severe
drought in in that year engulfed the entire Age of household head (yrs) 41 44 >0.50
No. of male adults 2 3 >0.50
country. The lowlands experienced the most No. of female adults 2 2 >0.50
severe drought and this may have contributed to No. of children 4 4 >0.50
its drop in maize area. Educ. household head (yrs) 6.5 5.8 >0.50
a Paired test comparing zones.
Comparison of farm size and maize area between zones show that the intermediate zone farms have
larger areas than lowland farms (Figures 3 and 4). The difference is not significant at p = 0.05. It
was expected that lowland households would have larger tracts of land than those in the intermediate
zone (Cunard et al. 1983; Nkonya et al. 1991; Nkonya et al. 1992). The reason for these findings is
that three of the four sample villages for the intermediate zone were drawn from Arusha districts,
namely, Hanang, Babati, and Karatu; districts known to have larger tracts of land than Arumeru and
Rombo, the districts representing the lowland villages (Nkonya et al. 1991).
Farmers were asked to state their future plans for maize area: fifty-three percent of the sample
farmers intended to keep their maize area constant, while 34% intended to increase it. Sample
farmers who planned to keep their maize area constant said that the area they had was enough for
their current food and marketing requirements. At the time of survey, farmers probably did not have
an incentive to grow more maize the following season because of low maize prices in 1995. The
average producer price of maize in 1995 was Tsh 4,500/100 kg bag (equivalent to US$ 72.5/t).
The average f.o.b price of white maize from South Africa was US$ 100/ton, i.e., US$ 10/100 kg
(Kapaliswa et al. 1992). Farmers who intended to increase maize area said they wanted to do so to
get more food and money.
3.3. Livestock ownership
Livestock keeping in the zone is quite common. Of the 126 farmers sampled, 84% kept cattle, 79%
kept goats, and 40% raised sheep. Herd size of all domestic stock was larger in Arusha than in
Kilimanjaro region. Overall, every household had about five head of cattle, six goats, and two sheep.
On the slopes of Mt. Kilimanjaro and Meru, farmers keep zero-grazed cattle, mainly for milk. Table
4a shows that the livestock numbers derived from sample farmers do not differ at a statistically
significant level across zones. Contrary to expectation, farmers in the intermediate zone tended to
have larger herds than those in the lowland zone. This may be the result of sampling more farmers
from Arusha for the intermediate zone than from Kilimanjaro. If the zones of the two regions were
sampled one at a time, it would be observed that herds of livestock in the lowland zone are bigger
than herds in the intermediate and highland zones. Hence, these results are due to a sampling
2.20- 4.5 0
Farm size ,
2.00- 4.00- Farm size
s 1.60. ----------------------- I gQQ__
- 1.60 3.00-
S1.040"- < 3 2.50- ..o '
1.20 "e 2.00 -
Maize area Maize area
1.00 1.50 i
1974 1984 1994 1974 1984 1994
Figure 3. Farm size and maize area in the lowland zone. Figure 4. Farm size and maize area in the intermediate
An examination of livestock ownership by method of land preparation shows that farmers who used
oxen had significantly bigger herds of cattle (p = 0.01) than hand-hoe users (Table 4b). Tractor users
also had significantly bigger herds of cattle than hand-hoe users (p = 0.10). Herds of small ruminants
did not differ significantly (p = 0.10) by method of land preparation or zone.
3.4 Farm mechanization
All sampled households had hoes equivalent to
the number of available family laborers. That is,
if a family had two people working on-farm, then
that household would have two hoes. Almost all
families sampled also had cutting tools such as
machetes, axes, and knives for different farm
Table 5 presents the number of ox-ploughs, ox-
carts, trailers, and tractors owned and used by
sample farmers in the two zones. Ox-ploughs are
used for land preparation and planting in
Arumeru, Babati, Hanang, and Mbulu districts in
Arusha region and Hai District in the Kilimanjaro
region. Comparison of ownership across zones
shows that there is no significant difference
between the mean number of implements owned
by sample farmers (p = 0.05). However,
comparing the ownership of farm implements
and machinery across the method of land
preparation shows a very distinct difference (p =
0.01). As expected farmers who use hand-hoes
for land preparation have the lowest mean
number of ox-ploughs, carts/trailers, and
tractors. Their decision to prepare land by hand
may be dictated by the terrain or by lack of
resources to buy or hire farm machinery. The
rolling topography in the highlands makes
mechanization difficult, forcing farmers to
prepare land by hand.
For the intermediate zone, the major means of
land preparation is hiring a tractor from rich
farmers (Table 6). A higher proportion of
farmers (p=0.05) in the intermediate zone used a
tractor for land preparation than farmers in the
Table 4a. Livestock ownership on zonal basis
Mean of: Lowland Intermediate P-valuea
No. of cattle 3 6 >0.50
No. of goat 6 6 1.0
No. of sheep 1 2 >0.50
No. other livestock 5 8 >0.50
a Paired test comparing zones.
Table 4b. Livestock herd size by land preparation
Cattle Goats Sheep
Land prep. method Mean heads -
Hand-hoe (n=27) 2.7 4.9 1.5
Oxen (n=24) 6.2 4.7 1.8
Tractor (n=54) 4.5 7.1 2.2
P-valuea: Hand-hoe vs. oxen 0.01 0.80 0.65
Hand-hoe vs. tractor 0.07 0.18 0.33
Oxen vs. tractor 0.80 0.57 0.89
a Paired test comparing methods of land preparation.
Table 5. Number of farm implements owned
Mean of: Lowland Intermediate P-valuea
No. of ox-ploughs 0.50 0.66 0.15
No. of carts 0.05 0.49 0.42
No. of tractors 0.05 0.37 0.34
Mean of: Hand-hoe Oxen Tractor P-valuea
Mean No.owned Hh/Ox Hh/Tra Ox/Trac
No. ofox-plough 0.12 1.08 0.61 0.001 0.00 0.00
No. of carts 0.04 0.87 0.40 0.00 0.00 0.00
No. of tractors 0.04 0.60 0.30 0.00 0.00 0.00
a Paired test comparing zones and methods of land
Table 6. Method of land preparation in each zone
Method Proportion P-valuea
Zero tillage 0.023
a Paired test comparing zones.
lowland zone. There were five tractors for each 100 sample farmers in the lowland zone as
compared to 37 tractors per 100 respondents in the intermediate zone (Table 5). The majority of
farmers owning tractors were found in Arusha. The major reason for this may be that Arusha farmers
residing in the high plateaus of Hanang, Babati, Mbulu, and Monduli have bigger tracts of land that
justify owning a tractor. Farmers in Kilimanjaro with small parcels of land have little incentive to own
tractors-they would rather hire than own tractors. Oxen are the second most important means of
land preparation in all zones. Land preparation by oxen is mainly used in the lowland and
intermediate zones. The highland zone has difficult terrain that hardly allows mechanization. As
opposed to a priori expectations, Table 6 shows that the lowland zone has a significantly higher
proportion (36%) of farmers using hand-hoe for land preparation than farmers in the intermediate
zone. The difference was significant (p = 0.01). Most hand-hoe users were sampled from Shimbi
village (Rombo), which is not typical of the lowlands. Other studies show that the use of hand-hoes is
common in the high rainfall, high altitude zone (Cunard et al. 1985; Nkonya et al. 1991). The high
rainfall, high altitude zone was not adequately sampled in this study due to its low area and
production of maize. The proportion of farmers using oxen in both zones was not significantly
4.0 MAIZE PRODUCTION PRACTICES AND
ADOPTION OF RECOMMENDATIONS
4.1 Crops and cropping systems
Intercropping maize with beans or pigeon peas is the most common cropping system in northern
Tanzania. Sixty percent of sample farmers in the intermediate zone grew maize in association with
beans and pigeon peas (Table 7). Pigeon peas are usually planted in Babati and the Karatu area in
Mbulu. It is an up-and-coming crop in Arumeru. Pigeon pea is considered a commercial crop as less
than 10% of production is consumed at home. Beans are widely grown in all districts growing maize.
As expected, more sample farmers in the lowlands grew maize in pure stands (40%) than in the
intermediate zone (Table 7). Moisture stress may be one reason for planting maize in pure stand.
Another reason may be that farmers in the lowlands do not have an acute land shortage, making
monocropping possible. The major reason cited by farmers for intercropping maize with legumes
was land scarcity. About 30% of farmers intercropped because of land shortage. Eighteen percent
said that they intercropped maize with legumes to get more money, while only 6% intercropped to
diversify production and spread price risks. Only 3% reported that they intercropped to save labor.
One of the most compelling reasons for intercropping is that it spreads risk. Many farmers did not
cite this reason because they regard production risks as beyond their control.
Coffee in association with banana is a common cropping system on the mountain slopes where
rainfall is high and temperatures are cool. About 19% of the sample farmers grew coffee in
association with banana. Tuber crops were not commonly grown in the zone. Only 1% of sample
farmers grew tuber crops, i.e., cassava or sweet potatoes.
4.2 Cropping calendar
The Northern Zone has a diverse climate ranging from temperate zones on mountain slopes to semi-
deserts in the lowland plains and the Maasai plateau. In most areas, the cropping season begins in
October and ends in September. Rainfall is generally bimodal (Figure 1). Short rains (vuli) are not
utilized for maize production in most areas. Rombo, Pare, and the Mwanga eastern highlands are on
the windward side of the Mt Kilimanjaro and Pare mountain ranges, however, and they capture
moisture-rich air from the ocean and use these vuli rains for maize production. For instance, the
Rombo season commences in July and ends in April. Figure 5 shows the cropping calendar for the
Table 7. Cropping systems in the two zones
About 21% of sample farmers in the
intermediate zone prepare their land in
November, 32% in December, and 29% in
January. The commencement of land
preparation in the lowland zone is more variable
as it captures two major seasons: the long season
Maizellegume Sole maize
Intermediate (n=90) 85.2 14.8
Lowland (n = 36) 60.0 40.0
P-valuea 0.00 0.04
a Paired test comparing zones.
in Arumeru (Mlangarini village), and the short season in Rombo District (Shimbi village). Fifty-three
percent of the sample farmers prepared their land between July and September. Most of the farmers
who prepared their land during this period were sampled from Shimbi Rombo District. It should be
noted that although the vuli season rain is not commonly used for maize production in the typical
lowland zone, Rombo is an atypical case. In the highland and intermediate zones, the vuli season is
normally used for vegetable production. Lowland farmers use irrigation for crop production in the
Farmers were asked why they chose to prepare land and plant when they did. More than 95% of
sample farmers in both zones said that the onset of rains determined their cropping calendar.
Farmers start land preparation well before the onset of rains. Both dry and wet planting are
common. The most common practice is to plant after effective rains have fallen because farmers
Figure 5. Maize cropping calendar for agroecological zones in northern Tanzania.
are then sure that there is enough moisture in the soil. The cropping season begins in Rombo, then
moves to Arusha and finishes in Moshi and Hai. The labor bottleneck occurs during planting and
weeding. So for Arusha, the labor bottleneck occurs in January through March while in Moshia the
labor shortage is experienced in February through April.
4.3 Rate and intensity of adoption of improved maize varieties
Feder et al. (1985) defined adoption as the degree of use of a new innovation in long-run equilibrium
when a farmer has full information about the new technology and its potential. If the innovation is
modified periodically, however, the equilibrium level of adoption will not be achieved. As the new
technology is introduced, some farmers will experiment with it before adopting. The rate of adoption
is defined as the percentage of farmers who have adopted a new technology or the area under a new
technology. The intensity of adoption is defined as the level of adoption of a given technology. For
instance, the number of hectares planted with improved seed or receiving fertilizer are measures of
the intensity of adoption of improved maize seed.
4.3.2 Current varieties grown
According to Table 8, all sample farmers grew locally bred or imported cultivars. Some farmers
maintained that they grew 'local' varieties, which they did not have local names for. Actually these
'local' varieties were improved varieties that have been recycled for so many years that they have lost
Some farmers in the eastern zone of Mbulu District, however, grew local maize varieties called 'bhor,'
which means 'black.' The bhor variety is flint, with mixed white and black grains, tall, and late
maturing (Nkonya et al. 1992). CG4141 is the most common variety grown in both zones and with
various methods of land preparation. With the exception of Katumani, there is an insignificant
difference (p = 0.05) in the proportion of farmers planting other varieties between the two zones.
Among the methods of land preparation, the proportion of farmers who used ox-plough planted
significantly (p =0.05) more H632/H622 than the respondents who used other means of land
preparation. H632/H622 are mostly planted in Babati, Mbulu, and Hanang districts, which use ox-
plough for land preparation and planting. Use of CG4141 and all composites was not significantly
different (p = 0.05) among farmers classified by method of land preparation (Table 12).
4.3.3. Trend of adoption of maize varieties
Figures 6 and 7 show the trend of adoption of all improved maize varieties for the past 20 years.
Comparison of adoption across zones indicates that there was no significant difference (p=0.05) in
trends of adoption over all the years. It was expected that lowland farmers would lag behind the
intermediate zone farmers but this was not the case. Figure 6 shows that before 1992, farmers in the
intermediate zone had adopted improved maize varieties more than those in the lowlands. After
1992, the adoption rate of lowland farmers surpassed the intermediate zone level. However, the
difference in the rate of adoption between the two zones continued to be insignificant (p = 0.05).
Comparison of the rate of adoption across the method of land preparation shows that hand-hoe users
lagged behind oxen and tractor users between 1985 and 1992. After 1992, hand-hoe users caught
up with the oxen and tractor users. The adoption patterns of oxen and tractor users tended to move
together throughout (Figure 7). The paired test shows that the difference in the rate of adoption by
method of land preparation is not significant (p = 0.05) for all the years and between any two
methods, except hand-hoe versus oxen in 1990, and hand-hoe versus tractor in 1990 and 1991.
As more farmers adopted improved varieties, maize yield increased over time (Figures 8 and 9). The
yield increase probably resulted from farmers gaining experience in management of the new
technology. The yield trend for 'local varieties' does not show pronounced changes for either zone or
methods of land preparation. This implies that those varieties have attained their maximum potential.
4.4 Factors affecting adoption of improved maize varieties
4.4.1 Household characteristics
Researchers and extensionists want to know the factors that affect both the rate and intensity of
adoption of new technologies. To a researcher, this is important for designing technologies that
40. zone .
S i zone
1974 80 85 90 91 92 93 94
Figure 6. Adoption of maize varieties by agroecological
1974 80 85 90 91 92 93 94
Hand-hoe (local) A Oxen (local) A Tractor (local)
OHand-hoe (improved) [ Oxen (improved) 0 Tractor (improved)
Figure 8. Local and improved maize yields by land
1974 80 85 90 91 92 93 94
Figure 7. Adoption of maize varieties by land preparation
1974 80 85 90 91 92 93 94
o Lowland (local var) o Lowland (improved var)
Intermediate (local var) Intermediate (improved)
Figure 9. Local and improved maize yield by zone.
address the factors that influence adoption. For an extension officer the knowledge helps in designing
extension messages that respond to the identified factors. To check for an annual variance of
intensity of adoption, a three-year average of acres under improved maize seed or fertilizer for each
sample farmer was used instead of a single year level of adoption. The factors that were thought to
influence adoption of both improved maize seed and fertilizer were: farm size, farming experience,
education of household head, livestock units, family labor, methods of land preparation (hand-hoe,
ox-plough, tractor, and other), rate of nitrogen application, and area under improved maize seed.
A two-step Heckman's procedure was used to analyze the adoption of inorganic fertilizer and
improved maize seed. In the first step of Heckman's procedure, the coefficients of factors that
influence the probability to adopt a technology (DP/dx) are estimated using a probit model. In the
second stage, the impacts of the same factors on intensity of adoption (DADIS/ax and DNRATE/dx)
are estimated using a system of Seemingly Unrelated Regression (SUR) equations. In this study, two
equations, one for ADIS and another for NRATE as dependent variables, are estimated
simultaneously in a system of equations. Tables 8 and 9 show the Heckman's first stage results for
the two zones separately and jointly.
In the lowland zone, 4 out of the 36 sample farmers did not use improved maize seed compared to
7 out of the 90 sample farmers in the intermediate zone. This high rate of adoption in both zones
reduced the variability of the dependent variable, making the influence of almost all regressors not
significant (p=0.05). As Table 8 shows, there was no factor that significantly affected (p=0.05) the
probability to adopt improved maize seed in the lowland zone. The same observation was made
when both zones were considered jointly (Table 9).
Table 8. Heckman's first stage procedure results estimating factors affecting adoption of improved maize seed for
the intermediate and lowland zones
Intermediate zone (n = 90) Lowland zone (n = 36)
Variable MLE dP/dx P-value MLE dP/dx P-value
Constant 3.79 0.550 >0.50 2.21 0.42 >0.50
Farm Size -0.01 -0.001 >0.50 -0.06 -0.01 0.25
Farming Experience 0.09 0.013 0.05 0.11 0.02 0.17
Education of HH1 Head 0.16 0.023 0.25 0.30 0.06 0.49
Livestock Units2 -0.03 -0.004 0.38 -0.03 -0.01 >0.50
Family Labor 0.35 0.052 0.13 0.47 0.09 0.25
MLP3: Hand-hoe -7.57 -1.105 >0.50 -7.70 -1.46 >0.50
Ox-Plough -5.19 -0.757 >0.50 -5.11 -0.97 >0.50
Tractor -5.58 -0.815 >0.50 -5.20 -0.98 >0.50
Nitrogen Fertilizer rate 0.01 0.001 >0.50 0.13 0.03 0.22
LLF = -16.43 cdf = 0.92, pdf 0.146 LRT4 (df= 9) = 16.34 LLF = -7.88 cdf= 0.89, pdf= .189 LRT (df=9) = 9.36
McFadden R2 = 0.33 McFadden R2= 0.37 LRT=9.36
X2 (df=9,p=0.05) =16.92
1 Household 2 Livestock Units: Cow = 0.8; Goat = 0.4; Sheep = 0.4
3 MLP = Method of Land Preparation. 4 LRT = Likelihood Ratio Test
The problem of analyzing factors affecting adoption in the lowland zone is that since almost all
farmers recycled improved maize seed, it was not easy to sample farmers who used 'local' varieties.
In a study by Nkonya et al. (1997), improved maize seed was defined as certified seed planted by
sample farmers. Hence, farmers who recycled seed were not counted as adopters, so the rate of
adoption was 52% as opposed to the 94% observed in this study. The factors that significantly
affected adoption of improved maize seed in the study by Nkonya et al. (1997) were level of
education of household head, farm size, and number of extension visits. This implies that future
adoption studies in northern Tanzania should distinguish between improved variety seeds that are
certified and those are not certified (recycled).
In the intermediate zone, however, farming experience positively influenced the probability to adopt
improved maize seed (Table 8). An increase in one year of experience increased the probability to
adopt (DP/dx) by 1.3%. If farming experience entails accumulation of knowledge, then it augments
technology use, as observed in the intermediate zone. Experienced farmers must have had the
opportunity to experiment with new varieties and observe their superiority over local varieties. This
argument is supported by Figure 8, which shows upward yield trends of improved maize over time.
However, if experience is equated with the aging process, it may then have a negative impact on
technology adoption as old farmers are set in their ways and tend to stick to old technologies. The
two antagonistic influences might have acted simultaneously, with the technology augmenting force
prevailing in the intermediate zone. Overall, the three models (summarized in Tables 8 and 9) for
estimating factors influencing probability to adopt improved maize seed are not significant (p = 0.05).
Table 10 reports the Heckman's second stage SUR system results for the intermediate zone and for
all zones. The SUR system for the lowland zone was not estimated because a very small sample
Table 9. Heckman's first stage procedure results estimating factors affecting adoption of improved maize seed and
fertilizer for the two zones
Intermediate zone (n = 90) Lowland zone (n = 36)
Variable MLE dP/dx P-value MLE 1P/3x P-value
Constant 5.201 0.590 >0.50 0.153 0.043 0.85
Farm Size -0.012 -0.001 >0.50 0.007 0.002 0.71
Farming Experience 0.033 0.004 0.18 0.010 0.003 0.40
Education of HH1 Head 0.104 0.012 0.31 0.021 0.006 0.72
Livestock Units2 -0.018 -0.002 0.47 -0.054 -0.015 0.02
Family Labor 0.158 0.018 0.26 0.024 0.007 0.54
MLP3: Hand-hoe -5.828 -0.650 >0.50 -0.350 -0.098 0.19
Ox-Plough -4.899 -0.554 >0.50 -0.741 -0.206 0.90
Tractor -5.211 -0.589 >0.50 -0.068 -0.019 0.53
Nitrogen fertilizer rate 0.005 0.001 >0.50 0.258 0.072 0.85
LLF = -22.75 cdf= 0.944, pdf= 0.113 LRT4 (df= 9)= 8.58 LLF= -77.34 cdf 0.587, pdf =0.278 LRT (df=9) = 16.13
McFadden R2 = 0.159 McFadden R2= 0.09 LRT=16.13
X2 (df=9,p=0.05) =16.92
1 Household 2 Livestock Units: Cow = 0.8; Goat = 0.4; Sheep = 0.4
3 MLP = Method of Land Preparation. 4 LRT = Likelihood Ratio Test
remained after omitting the non-adopters of both technologies (only 11 observations were adopters
of both improved maize seed and fertilizer). Consequently, the system ran out of degrees of freedom
because the total number of independent variables (10 for ADIS and 10 for NRATE) was more than
the number of remaining observations. In the intermediate zone, no independent variable (other than
the constant) significantly affected intensity of adoption of improved maize seed (p = 0.05). The
reason for the non-significant results may be the narrow range of intensity of adoption (0.17 to 1.0
ha, with a mean of 0.89 and a standard deviation of 0.21 ha). The low variability may be due to the
use of the three-year average, which tended to smooth over the variability of adoption among
With the lack of significance of the independent variables, we need to look at other sample and non-
sample variables that influence intensity of adoption. Such variables could not be included in the
Heckman's procedure because they were either non-sample information (e.g., seed marketing
problems and maize seed variety characteristics), or they were farmers' responses/observations when
asked why they adopted/disadopted technologies and/or why they preferred some varieties over
others. The discussion of market and institutional factors affecting adoption of maize technologies
and reasons for varietal preference and disadoption follows in the next section.
Table 10. Heckman's second stage procedure SUR results estimating factors affecting intensity of adoption of
improved maize seed and fertilizer for intermediate zone and for the two zones
Intermediate zone Both zones
Variable Coefficients Std error P-value Coefficients Std error P-value
First equation of SUR (intensity of adoption of improved maize seed)
Constant 0.674 0.212 0.00 0.836 0.197 0.00
Farm size -0.001 0.004 >0.50 0.002 0.004 >0.50
Farming experience 0.002 0.004 >0.50 0.001 0.004 >0.50
Education of HH head 0.002 0.016 0.15 0.007 0.015 >0.50
Livestock units -0.006 0.008 0.47 -0.003 0.007 >0.50
Family labor 0.009 0.011 0.42 -0.001 0.010 >0.50
MLP: hand-hoe -0.019 0.180 >0.50 0.029 0.158 >0.50
Ox-plough 0.008 0.135 >0.50 0.030 0.127 >0.50
Tractor -0.064 0.117 >0.50 -0.016 0.116 >0.50
Nitrogen fertilizer rate 0.001 0.001 0.40 0.001 0.001 0.34
IMRS1 0.431 0.374 0.25 -0.256 0.472 >0.50
Second equation of SUR (intensity of adoption of fertilizer)
Constant 194.76 125.10 0.13 73.61 137.10 >0.50
Farm size -0.847 0.662 0.20 -1.038 0.992 0.29
Farming experience -2.778 1.459 0.06 -1.141 1.162 0.33
Education of HH head -1.506 2.937 >0.50 -0.566 3.338 >0.50
Livestock units 9.817 6.018 0.10 4.887 6.131 0.42
Family labor -0.771 1.976 >0.50 -0.390 3.273 >0.50
MLP: hand-hoe 39.998 22.663 0.08 46.427 40.148 0.25
Ox-plough 29.341 33.316 0.38 50.362 87.399 >0.50
Tractor 17.296 18.149 0.34 25.311 19.116 0.19
Nitrogen fertilizer rate -56.561 50.721 0.27 13.524 31.831 >0.50
IMRF2 -216.96 144.62 0.13 -122.390 190.080 >0.50
System R2 = 0.283; Observations above limit (OAL)= 54 System R2 = 0.178 OAL=71
1 Inverse Mills Ratio for seed.
2 Inverse Mills Ratio for fertilizer.
4.4.2 Market and institutional factors
Farmers reported that the major factors affecting adoption of improved maize varieties are input
prices, the marketing system, and varietal traits. The discussion below examines the maize seed
industry in northern Tanzania and varietal traits that farmers prefer or dislike that affect adoption/
disadoption of some varieties.
4.4.3 The maize seed industry in Tanzania
CG4141 is multiplied and distributed by Cargill Hybrid Seed Ltd. based in Arusha. The locally bred
hybrids H622 and H632 are grown mainly by farmers in the intermediate zone. Only 6% of sample
farmers in the lowland grew the locally bred hybrids. This is because the hybrids are late maturing.
Locally bred cultivars are flint with good pounding and storage qualities, and as high yielding as
CG4141. They are marketed mainly by the Tanzania Seed Company (TANSEED) which has not fared
well in the competitive seed industry. This has contributed to the low adoption of locally bred hybrids.
Before input liberalization in 1990, locally bred varieties were nearly the only improved maize planted
in Tanzania. Now, the demand for (adoption of) locally bred cultivars in northern Tanzania is low.
After input liberalization private companies have engaged in seed multiplication and also conducted
trials that evaluate the adaptability of imported varieties to the local environment. The varieties are
subsequently released to farmers. CG4141 is competing aggressively with the locally bred cultivars
that are multiplied and sold by TANSEED. Pannar, meanwhile, started production and marketing of
maize seed in 1995. The new companies recruit a chain of stockists who sell their seed in villages and
towns. TANSEED has followed suit. Farmers also indicate that seed sold by the private companies is
purer, more uniform, and higher yielding than TANSEED seed. This has reduced the demand for
The biggest drawbacks of the new varieties sold by Cargill and Pannar are their high price, poor
storability, pounding quality, and taste. Pounded maize is used to make a local dish prepared from
grains whose seedcoat has been removed (kande). Some farmers also pound their maize before
milling to make a whiter and softer dough (ugali). When pounded, maize grain with a soft seedcoat
breaks and flour losses before milling are greater. This underscores the importance of the flint trait in
farmers' varietal preferences.
The latest development in the maize seed industry is the renewed importation by the Tanganyika
Farmers Association (TFA) of the once-famous hybrid H511 from Kenya. H511 is a flint maize and is
as high yielding and as early maturing as CG4141. Its advantage over CG4141 is its flinty grain.
The 1994/95 price for Cargill (CG4141) and Pannar (PAN 6481) seed was 650 Tanzanian shillings
(Tsh) /kg, while Kilima, a composite, sold at Tsh 450/kg. The high prices of maize seed have forced
many farmers to recycle hybrids. No sample farmer reported growing Staha, TMV1, or Tuxpefo.
Demand for composites is generally low, despite their possessing a distinct advantage: with proper
selection in the field, composites can be recycled for at least three years without substantial loss of
vigor. This concerns the local breeders who have strived to produce composite varieties for the
Northern Zone (Moshi et al. 1990).
It is not clear whether the low demand for composite varieties is due to a low standard of seed
multiplication resulting in poor seed quality or farmers' lack of information about the advantage of
recycling OPVs with less yield loss than recycling hybrids. Marketing efforts by the TANSEED
company, which sells the locally bred cultivars, are much less than those of Cargill and other
companies selling imported seed. This obviously contributes to the low adoption of the local
improved seed. Another possible reason for the low demand for composites is that the seed industry
does not make substantial profits on OPVs, because farmers normally recycle OPV seed for three
years, during which time the seed dealers would experience low seed demand. This is the major
reason why Cargill Hybrid Seed Ltd. does not sell very much OPV seed (Banfield, Director Cargill
Hybrid (T) Ltd., personal communication). OPVs also have a low yield potential relative to hybrids
during good seasons, yet another possible explanation for the low demand for OPVs.
Before input liberalization, quasi-governmental institutions and cooperative unions monopolized input
marketing. These institutions were inefficient in delivering inputs to farmers. They suffered from
chronic liquidity problems since they depended on borrowed money for buying inputs. This led to
delayed input supply and chronic shortages that served as a disincentive to farmers (Mbiha 1993;
Nkonya 1994). Input liberalization has led to a rapid increase in the number of private businesses
that engage in input marketing. Farmers could obtain inputs, in a timely manner, from village
stockists who are much closer to them than prior to 1990. As expected, the price of inputs has
increased sharply, wiping out the shortages that existed before.
4.4.4 High price of inputs coupled with low price of maize produce
All sample farmers who recycled improved varieties said that the rising certified seed and fertilizer
prices coupled with falling maize output price was the major reason for recycling improved maize
seed. Table 11 shows the price ratios of maize produce to the price of fertilizer and maize certified
seed. The ratios give the amount of N (or certified maize seed) that 1 unit of maize produce can buy
given their market prices. This represents the actual purchasing power of the farmers.
[Tsh/(kg of Maize)]/[Tsh/(kg of N)] = (kg of N)/(kg of Maize) Table 11. Trend of maize produce price to
input price ratio in Tanzania
The purchasing power of maize produce has fallen from its Year N H632/H622 OPVs
highest level in 1984 to its lowest level in 1995. For 1980 0.003 0.123 0.167
instance, the purchasing power of maize produce fell from 1981 0.004 0.139 0.188
0.011 kg to 0.002 kg of N per kg of maize produce 1982 0.012 0.333 0.455
between 1984 and 1995. 1983 0.024 0.915 1.204
between 1984 and 1995.
1984 0.011 0.442 0.543
1985 0.009 0.282 0.372
Trade liberalization in Tanzania was initiated in 1984 and 1986 0.008 0.188 0.261
1987 0.007 0.137 0.169
this caused a considerable fall in the market price, hence 1988 0.008 0.106 0.147
the fall in terms of trade for maize farmers as compared 1989 0.010 0.133 0.152
to input dealers (Missiaen and Lindert 1993). The trend of 1990 0.010 0.109 0.122
1991 0.006 0.205 0.242
the purchasing power of maize produce in relation to 1992 0.006 0.207 0.229
maize seed stayed nearly constant between 1988 and 1993 0.006 0.108 0.144
1990, and it rose between 1990 to 1991 as this period 1994 0.004 0.072 0.102
1995 0.002 0.039 0.070
followed a poor harvest in 1990/91. It fell steadily between 1986 and 1995. Fertilizer price
increased rapidly during 1990-1994 when the government continued to withdraw its subsidy from
the peak level of more than 50% that prevailed in the early 1980's, to 0% in 1993 (MOA 1993b).
The 1994/95 price of Cargil maize seed varieties (CG4141) and Pannar (PAN 6481) was Tanzanian
shillings (Tsh) 650/kg, while Kilima, a composite, sold at Tsh 450/kg.
4.4.5 Varietal traits preferred by farmers
As suggested by the results of Table 12, Table 13 also shows that CG4141 was the most preferred
variety in both zones. However, the proportion of farmers preferring CG4141 and Katumani in the
lowlands is significantly higher (p=0.05) than the proportion in the intermediate zone. The probable
reason for this is that Katumani and CG4141 mature earlier than the locally bred hybrids, so lowland
farmers with less reliable rainfall would naturally choose the early maturing varieties. Farmers in the
intermediate zone could opt for varieties other than CG4141 and Katumani because the moisture
limitation is less acute than in the lowlands.
In both zones, high yield is the major reason for preferring varieties. CG4141 again was identified as
the most high yielding in both zones. Drought tolerance/resistance/avoidance was the second major
criterion for preferring a variety and CG4141 once again was selected for that reason (Table 14).
Table 12. Maize varieties planted in the 1994/95 season
Intermediate Lowland Hand-hoe Oxen Tractor
zone (n=88) zone (n=18) (n=26) (n=28) (n=55) P-valuesb
Variety % Plantingc P- valuea % Plantingc Hh/Ox Hh/T Ox/T
CG4141 55.7 61.1 >0.50 73.0 64.0 80.0 0.50 0.10 0.10
H632/H622 23.9 5.6 0.01 8.0 29.0 9.0 0.02 0.01 0.02
Kilima 16.0 28 0.29 19.0 7.0 7.0 0.20 >0.50 >0.50
UCA 3.4 0.0 0.08 1.0 1.0 1.0 1.00 1.00 1.00
Katumani 0.0 5.6 0.30 0.02 0.04 0.0 >0.50 0.20 0.15
Staha 1.1 0.0 0.32 1.0 0.0 0.0 >0.50 1.00 1.00
a Paired test comparing zones.
b Paired test comparing methods of land preparation. Each two methods of land preparation are compared using a paired t-test.
The results are reported in the last three columns of Table 8. The pairs of land preparation compared are: Hh/Ox = hand-hoe/
oxen; Hh/T = hand-hoeltractor; Ox/T = oxen/tractor.
c Percentages may add to more than 100% because some sample farmers plant more than one variety and/or use more than one
method of land preparation.
Table 13. Most preferred maize varieties
Table 14. Reasons for varietal preference
CG4141 54.2 77.8
H622/632 22.9 0.0
Kilima 16.0 6.0
Katumani 0.0 17.0
a Paired test comparing the two zones.
Intermediate (n=83) CG4141
Lowland (n=36) CG4141
% Reporting reason
yield Drought Sweet Other
31.3 20.5 2.4 0.0
18.1 3.6 0.0 3.6
4.8 6.0 3.6 0.0
52.8 19.4 0.0 5.6
0.0 5.6 5.6 5.6
5.6 0.0 0.0 0.0
4.4.6 Factors contributing to disadoption of improved varieties
Forty-five percent of sample farmers had at one point discontinued growing an improved variety for
various reasons. The varieties discontinued by most farmers in the intermediate zone were hybrids
H622 and H632, and in the lowland zone, Kilima and Katumani (Table 15). The major reason given
for discontinuing hybrids was their late maturity. Given that most maize is produced in the
intermediate and lowland zones, moisture may be limiting in some poor years, so farmers would
prefer medium to short season varieties. H632 and H622 reach physiological maturity after 150-160
days. CG4141 and other imported varieties were disadopted by 28%, mainly because they were no
longer available. Importation of SR52 variety from Zimbabwe was discontinued in 1988. This was
also the case for the MH41 (CG4141) that was not available until Cargill Hybrid Seed Ltd. started
multiplying and distributing the variety in 1990. Other varieties were disadopted for being low
4.5 Seedbed type, planting configuration, and weeding
4.5.1 Seedbed type
The common seedbed type in northern Tanzania in both zones is the flat bed (Table 16). All farmers
said that they use a flat seedbed because it is easy to work. Seedbed types are important for water
harvesting in semi-arid areas. For instance, tie ridging has been shown to conserve more moisture
than a flat bed. However, seedbeds are also determined by the method of land preparation. Most of
the tractors owned by medium-scale farmers in northern Tanzania do not have ridgers. This forces
farmers who hire them to prepare a flat bed.
Row planting was practiced by 99% of respondents. Only one farmer (from Rombo) out of 126
respondents broadcast seed. The major reason given for row planting is that it is easy to work. The
average row-to-row and hill-to-hill spacings are shown in Table 17. The average plant population as
reported by farmers was slightly lower than the
recommended population of 44,444 plants/ha Table 16. Percentage of farmers planting on flat beds or
for the medium and long maturing varieties ridges in each zone
commonly grown in the zone. Farmers in Zone Flat Ridges
Arusha who normally plant using ox-plough Intermediate (n=88) 100 0.0
usually attain the recommended level. Lowland (n=36) 97 2.8
Table 15. Discontinued varieties
% % P-valuesb
Intermediate Lowland Hand-hoe Oxen Tractor
Variety n=43 n=8 P-valuea n=10 n=11 n=31 Hh/O HhlT O/T
H622/632 48.8 12.5 0.01 20.0 37.0 48.0 0.40 0.50 >0.50
CG4141 25.6 12.5 0.33 30.0 9.0 35.0 0.25 0.10 0.10
Kilima 9.4 25.0 0.33 0.0 27.0 1.0 0.05 0.08 0.15
Katumani 2.3 25.0 0.14 20.0 9.0 3.0 0.50 0.40 0.40
a Paired test comparing the two zones.
b Paired test comparing methods of land preparation methods
Sometimes farmers exceed the recommended population because they drop seed in furrows made by
the ox-plough during harrowing. Occasionally farmers drop too much seed and later thin (Nkonya et
al. 1991). This is always the case when the seed is the progeny of improved varieties. When fresh
improved seed is used farmers tend to stick to the recommendation.
The plant spacing for monocropped maize and maize in association with legumes does not differ for
cases where maize is the major crop in the intercrop. However, farmers in Arumeru, Moshi, Rombo,
Babati, and Hai plant beans in association with maize in strips. In this case, bean is the major crop
and maize is a companion crop; more than two rows of beans are planted between two rows of
Table 18 shows that the majority of farmers weed twice. In the intermediate zone, 89% of
respondents weeded twice, compared to 60% of respondents in the lowlands. The difference of
proportion of farmers weeding twice between the two zones is significant (p=0.01). The difference in
the frequency of weeding between the two zones may be determined by the length of the growing
season. In the lowlands, the growing season is shorter, hence farmers may not have time to weed
twice. Only 21% of the sample farmers in the lowlands weeded their maize fields once (Table 18).
Table 19 shows the timing of planting, first weeding, and fertilizer application. On average, sample
farmers in the intermediate zone did their first weeding 3.8 weeks after planting compared to 3.6
weeks for the lowland farmers.
Table 17. Spacing for monocropped maize and maize/
Spacing (cm), Seeds/ Population Intermediate Lowland
rowx hill hill (plants/ha) (n=87) (n=22)
90 x 50 2 44,444 35.0 27.3
90 x 30 1 37,037 13.0 18.2
100 x 30 2 66,666 3.4 9.6
75 x 60 2 44,444 10.0 4.5
80 x 50 2 50,000 4.4 4.5
Broadcast 0.0 9.1
Other 39.7 27.6
Overall 1.8 44,408
Table 18. Frequency of weeding
Weeding (n=85) (n=33)
frequency % of farmers P-valuea
Once 3.5 21.2 0.02
Twice 89.0 63.6 0.00
Thrice 7.1 15.2 0.24
a Paired test comparing the two zones.
Assuming that emergence of planted maize
takes place a week after planting, the first
weeding is done two and a half weeks after
emergence. This is within the recommendation.
Matowo and Mgema (1990a) observed that
highest grain maize yield was obtained when the
first weeding was done 2-3 weeks after
Table 19. Time of planting, weeding, and fertilizer
Operation Mean weeks after planting P-valuea
Time of planting 9.5 (n=36) 6.9 (n=85) 0.00
Time of 1stweeding 3.8 (n=31) 3.6 (n=78) >0.50
2nd weeding 5.7 (n=18) 5.8 (n=32) >0.50
1st top dress 4.8 (n=10) 4.7 (n=36) >0.50
2nd top dress 5.8 (n=31) 5.5 (n=16) >0.50
a Paired test comparing the two zones.
b Weeks after land preparation.
emergence. On average, the second weeding is done 5.8 weeks after planting. This timing is within
the recommendation of performing both weedings within 6 weeks after emergence. Overall,
Northern Zone farmers closely followed the recommendation on weeding schedules. Herbicide use
was reported by only 7% of respondents.
Farmers reported that crop planting is dictated by the onset of rains. Therefore, the significant
difference in timing of planting (p=0.01) is a factor of location and the onset of rains, rather than
management differences. The number of weeks (after planting) to first and second weeding and
fertilizer application in both zones was not statistically different (p=0.05). This implies that after the
onset of the rains, farmers in both zones follow more or less the same cropping calendar.
4.6 Fertility management
4.6.1 Adoption of fertility management technologies
Table 20 shows that 80% and 92% of sample farmers in the intermediate and lowland zones,
respectively, apply some form of fertilizer. The difference in the percentage of farmers applying
some form of fertilizer between the two zones is significant (p=0.01). Sixty-four percent of
respondents in the intermediate zone applied chemical fertilizer compared to 44% of sample farmers
in the lowland zone in the 1994/95 season. The difference in the proportion applying chemical
fertilizer is significant at the 5% level. This result was anticipated because the intermediate zone
farmers receive more rain and their terrain is hilly, so leaching and erosion make their soils poor.
Adoption of chemical fertilizer is far less than adoption of improved maize varieties, which is almost
100%. This may be explained by the stepwise adoption of technologies, i.e., farmers decide to adopt
seed technology first because it is easily implemented, and adopt fertilizer later. Seed technologies are
normally adopted spontaneously while fertilizer requires a higher level of knowledge before farmers
decide to use it.
Application of organic manure is more common in the Arusha region than in Kilimanjaro. About
49% of sample farmers in the intermediate zone and 44% of respondents in the lowland zone
applied kraal manure (Table 20). Use of kraal manure and crop residue between the two zones
revealed no statistical differenences (p=0.05). These findings were unexpected, as it was anticipated
that farmers in the lowland zone would use more kraal manure and crop residue because they have
more cattle and they do not stall feed their
animals. The reason behind these findings may
Table 20. Adoption of fertilizer use in the 1994/95 season
be that the three villages sampled in the
Intermediate Lowland intermediate zone were from the livestock-rich
(n=86) (n=36) region of Arusha. This is reflected in Table 4a,
Type of fertilizer % of farmers applying P-valuea which shows that the intermediate zone farmers
had more cattle than those in the lowland zone.
Any form 80 92 0.00 Green manuring was uncommon as only 4% of
Chemical 64 44 0.04
Kraal manure 49 44 >0.50 sample farmers reported practicing it.
Crop residue 41 44 >0.50
a Paired test comparing the two zones.
4.6.2 Fallowing and crop rotation
Due to land shortage, only 3% of sample farmers in the intermediate zone fallowed their farms
compared to 13% in the lowland zone. This means that the most feasible method for replenishing
fertility in this zone is through the application of fertilizer. Crop rotation was reported by 1 1% of the
intermediate zone respondents and by 22% of sample farmers in the lowland zone. This may be
caused by land shortage that acts as a constraint on growing a variety of crops (Table 21).
4.6.3 Chemical fertilizer application trends
Figures 10 and 11 depict the trend in the
application of chemical fertilizer, indicating an
increase in the amount of N applied. The major
N-carriers used are SA, CAN, and urea. In both
zones, a dramatic increase of N fertilizer
application occurred after 1992. As was the case
with improved maize seed, this may be a
reflection of improved extension services and
input marketing. Maize yields have increased
accordingly as depicted by Figures 8 and 9. The
increase in yield may be due to adoption of
improved varieties and application of fertilizer.
With the exception of 1994, the intermediate
zone respondents applied more N fertilizer than
those in the lowland zone. The difference in the
amount of N applied between the two zones,
however, is not significant (p=0.05) for all years
studied. The lowland zone farmers increased N
application much faster from 1985 to 1994
when they caught up with those in the
intermediate zone. The reason for the dramatic
increase in N application in the lowland zone
may be that crop production started later than in
the intermediate zone (Dunford 1980). In the
past, fertility in the lowland zone may have been
higher than in the intermediate zone, hence the
farmers may have had good production without
applying much N fertilizer. Under continuous
cultivation, fertility in the marginal lowland zone
may have declined, thus requiring N applications
commensurate with those in the intermediate
Table 21. Fallowing and crop rotation
Fallow Crop rotation
Intermediate (n = 90) 3.1 13.2
Lowland (n = 36) 11.0 22.0
P-valuea 0.15 0.26
a Paired test comparing two zones.
1974 80 85 90 91 92 93 94
Figure 10. Fertilizer application by land preparation
1974 80 85 90 91 92 93 94
Figure 11. Fertilizer application by agroecological zone.
Evaluating the level of N applied by method of land preparation presents interesting results. Table 22
shows that, for the years of the study, there was no significant difference in the level of N applied
between farmers who used hand-hoes and those who used oxen for land preparation. There was also
no statistically significant difference in the level of N applied by farmers using hand-hoe and those
using tractors. A highly significant difference was found (p=0.01) in the level of N applied by farmers
who used tractors and those who used oxen. Farmers using oxen applied the lowest amount of N. It
was expected that farmers who used hand-hoes were the poorest and so they were expected to apply
the lowest level of N. Farmers who used oxen applied the lowest amount of N, probably because
they applied more kraal manure to their fields. The amount of kraal manure applied may be
determined by the number of livestock owned (Nyaki et al. 1991). Farmers who used oxen for land
preparation had a significantly higher number of cattle (Table 4b) (p = 0.07) than the tractor and
hand-hoe users, hence they were likely to apply more kraal manure than hand-hoe users. Farmers
who use tractors may be wealthier and more able to afford chemical fertilizer. They may also have
smaller herds of cattle that cannot supply enough kraal manure.
Until 1994, N application levels were still less than half of the recommendation for intermediate and
highland areas, which average 80 kg N/ha. The recommendation for the high rainfall area with an
altitude less than 900 masl was to apply 20-45 kg N/ha (Mowo et al. 1992; Samki et al. 1984).
Since farmers applied an average of 27 kg N/ha in 1994 (Figure 10), a large gap was evident
between recommendation and farmer practice. The majority of the farmers reported that they top
dress fertilizer only once. Only 10% of sample farmers applied a second top dressing of fertilizer,
while only 4% applied basal chemical fertilizer at planting.
4.6.4 Timing of fertilizer application
Timing of fertilizer application was close to the recommendation. There was no significant difference
(p=0.05) between zones in timing of fertilizer application. The maximum and minimum number of
weeks the sample farmers top-dressed fertilizer after planting maize was 13 and 1, respectively. The
first top dress was at about 4.7 weeks after planting for both zones (Table 19). This was also within
the recommendation of top dressing when maize is about one meter high.
Table 22. Trend of fertilizer application
Intermediate Lowland Hand-hoe Oxen Tractor
(n=31) (n=14) (n=27) (n=29) (n=61) P-valuesb
Year Mean N/ha P-valuea Mean N/ha HhlO HhlT o/T
1974 0.32 0.0 0.52 0.0 0.0 0.5 1.00 0.50 0.00
1980 2.42 1.6 0.70 0.5 0.0 4.3 0.00 0.19 0.00
1985 8.87 1.9 0.10 3.1 1.4 12.2 0.83 0.12 0.00
1990 12.7 4.0 0.12 10.7 2.3 14.9 0.61 0.57 0.00
1991 16.9 6.2 0.08 15.2 2.7 19.5 0.34 0.60 0.00
1992 17.9 10.2 0.26 14.9 3.4 22.2 0.32 0.42 0.00
1993 22.1 15.0 0.32 25.4 5.4 26.2 0.14 0.93 0.00
1994 24.2 26.7 0.75 26.3 12.4 32.7 0.29 0.52 0.00
a Paired test comparing the two zones.
b Paired test comparing methods of land preparation for the two zones.
4.6.5 Methods of fertilizer application
The majority of respondents in the intermediate zone broadcast fertilizer, contrary to the
recommendation for spot or furrow application and covering fertilizer with soil. In the lowland zone,
the majority of farmers banded fertilizer around crops without covering (Table 23). This means the
recommendation for fertilizer placement has been poorly followed in the intermediate zone where
nutrient application is more important. The difference in the methods of placement between the two
zones is significantly different (p=0.05) for all the methods except furrow application, which was not
practiced by any of the sample farmers.
4.6.6 Factors affecting adoption of fertilizers
As was the case for adoption of improved maize seed, impacts of some variables on intensity of
adoption fertilizer were analyzed using the two-step Heckman's procedure. Table 13 summarizes the
influence of selected variables on probability of adopting fertilizer for the two zones. Only livestock
units significantly influenced the probability of adopting fertilizer (p = 0.05). A unit increase in LUNIT
decreases the probability of adopting fertilizer by 1.5%. It was expected that, with large herd size,
farmers have more wealth and can therefore afford to buy chemical fertilizer, so LUNIT was expected
show positive indicators. Nkonya et al. (1997) also observed that herd size was negatively related to
the probability of adopting chemical fertilizer. The reason for this unexpected result may be explained
by the fact that farmers with large herds have more manure that is used to replace chemical fertilizer.
It is also true that farmers who have large herds of cattle live in areas with lower rainfall that are less
subject to soil leaching. This keeps their soils more fertile and reduces the need for fertilizer.
The second part of Table 10 summarizes the influence of the selected variables on intensity of
adoption for the intermediate zone and for the two zones taken together. For the two zones, no
variable significantly influenced NRATE (p=0.05). For the intermediate zone, however, LUNIT
significantly influenced NRATE (p=0.1). LUNIT has a positive impact on the probability of adopting
fertilizer for the two zones.2 This means that among adopters, LUNIT exerts a positive influence on
NRATE via its income effect, i.e., richer farmers (as indicated by bigger LUNIT) were likely to apply
4.6.7 Cron residue management Table 23. Methods of placing top dress fertilizer
The recommendation for crop residue
management for farmers who apply only low
levels of fertilizer or none at all has has been to
plough under in order to avoid soil mining.
About 94% of the respondents from the lowland
zone compared to 71% from the intermediate
zone reported that they fed their maize stover to
cattle in situ, i.e., in the field (Table 24). The
method % of farmers P-valuea
Broadcast 74.2 28.6 0.00
Spot (hole) 16.1 0.0 0.01
Banding 3.1 71.4 0.00
Furrow 0.0 0.0 1.00
a Paired test comparing the two zones.
2 It should also be noted that the opposing results are not surprising as they were obtained from two separate sub
samples, one consisting of non-adopters and the other adopters only. The negative impact of LUNIT on probability
to adopt was on non-adopters and its positive impact was on the intensity of adoption among adopters.
majority of farmers in the lowland zone did not zero-graze because they had large herds of cattle; the
herds are fed on stover after harvest. Only 14% cut and carried stover home to feed zero-grazed
dairy cows, resulting in a nutrient flow from the field to other areas. Some districts have bylaws that
restrict feeding crop residues in situ. Restrictions on grazing stover in the field help minimize soil
erosion because such grazing leaves the fields
Table 24. Management of crop residues bare and the animals pulverize the soil, making it
Intermediate Lowland Hand-hoe Oxen Tractor susceptible to wind and sheet erosion. Extension
(n=70) (n=18) (n=26) (n=14) (n=44) officers should encourage farmers who cut and
Management Proportion practicing carry stover, and those who feed crop residues
in situ, to apply fertilizer to their fields in order
Plough under 0.13 0.00 0.12 0.00 0.10 to replenish the exported nutrients. Sample
Burn 0.014 0.00 0.00 0.00 0.02
Feed cattle 0.71 0.94 0.65 0.93 0.75 farmers who ploughed under crop residues were
Stallfeed 0.14 0.06 0.23 0.07 0.10 1 1% of all the respondents.
4.7 Pest and disease control
4.7.1 Most serious pests and diseases
Insect pests are a significant production constraint for maize in Tanzania. Stalk borer (Buseola spp.),
one of the nation's most serious pests (Nyambo and Kabissa 1990), is prevalent during dry spells or
in areas with hot weather and marginal rainfall. Cultural and chemical control measures of the pest
have been recommended (Nyambo and Kabissa 1990). As expected, a significantly higher
percentage of farmers in the lowland zone (with hotter climate) reported being affected by stalk
borers compared to respondents from the cooler intermediate zone (Table 25). About 58% of
farmers in the intermediate zone who reported being affected by insect pests used chemical control
compared to 20% in the lowland zone. Other farmers did not use any control measures, probably
because the economic threshold of damage was not attained.
Table 25. Common maize pests reported by sample
Stalk borer Armyworm Cutworms
Intermediate (n=88) 53.0 42.0 5.70
Lowland (n=34) 77.0 19.0 4.00
P-value b (Zones) 0.00 0.00 >0.50
Hand-hoe (n=25) 68.0 28.0 4.00
Oxen (n=25) 80.0 16.0 4.00
Tractor (n=54) 48.0 44.0 8.00
P-valueb: (Hh/Ox) 0.30 0.30 1.00
(Hh/Trac) 0.10 0.20 >0.50
(Ox/Trac) 0.01 0.01 >0.50
a MLP = method of land preparation.
b Paired test comparing the two zones and methods of land
Armyworms occur occasionally. At such times,
the government takes responsibility for
controlling the outbreaks through the
department of pest control. Occurrence of
cutworms is spotty; it was reported by only 5.7%
of the respondents in the intermediate zone and
4% of sample farmers in the lowland zone.
Thanks to varieties bred for disease resistance
and tolerance, maize diseases were rarely
reported by sample farmers. The major disease
reported was MSV. No farmer reported having
controlled the disease. Its control includes
roguing and seed dressing using carbofuran
(Mduruma et al. 1990). Only two farmers in the
lowland zone reported MSV as a problem.
4.8 Transportation, storage, and post-harvest technology
As shown in Table 26, head loads are the commonest means of transporting maize from the fields to
homes for both zones and for all methods of land preparation.
Ox-carts and trailers were the second most important means of transportation for the two zones and
methods of land preparation. Use of donkeys was not reported by farmers, but other studies have
reported that the animals are used by many farmers in Mbulu, Arumeru, Hanang, Babati, Simanjiro,
Monduli, and Kiteto (Nkonya et al. 1992; Cunard et al. 1983).
The major maize storage pests in northern Tanzania are Sitophilus spp., Prostephanus truncatus
(Larger grain borer, LGB), Tribolium spp., Rhizopertha dominica, ephestia spp., Sitotroga
cerealla, and Oryzaephilus spp. Storage losses of maize of up to 70% have been recorded in the
lowland zone as a result of LGB (Uronu 1990). Storage losses caused by other pests are substantial if
stored maize is not treated.
Farmers used different methods of preserving their maize crop (Table 27). Most farmers in
Kilimanjaro stored their maize using gunny bags and airtight drums. Upright cribs were common in
Arusha, where 60% of intermediate zone
respondents were sampled. Farmers who used Table 27. Methods of grain storage across zones
cribs normally did not shell their maize. Storage
Method Intermediate Lowland
is accomplished by tying two cobs of maize
together using the husks and hanging them on %
upright cribs. When farmers need to mill maize, 57
On cribsa 56.7 2.8
they shell the amount needed at that time. This Gunny bag 22.2 88.9
method allows farmers to forego treating their Air tightdrums 21.1 5.6
maize. Kihengeb 1.3 2.8
Shelled no treatment 14.1 2.8
Use of gunny bags is the most common practice
in the lowland zone (Table 27). Of sample
a Unshelled and untreated.
b Maize is shelled and treated with ash or other local material
and then stored in local containers called 'kihenge'.
Table 26. Transportation of maize from field to homestead
Intermediate Lowland Hand-hoe Oxen Tractor
(n= 70) (n = 18) (n = 26) (n=14) (n=44)
Head loads 42.5 57.6 55.0 39.0 48.0
Carts 41.4 21.2 22.0 44.0 33.0
Pick-up 16.1 12.1 11.0 17.0 19.0
Bicycles 0.0 9.1 11.0 0.0 0.0
farmers who used gunny bags, 80% treated maize using chemicals alone or in combination with
ashes. Farmers who used drums did not use any chemicals and reported that they did not experience
grain losses. Treatment with actellic super dust for shelled maize was the most common chemical
control. Ninety percent of tractor users treated their shelled maize compared to 77% of hand-hoe
4.9 Seed selection
4.9.1 Seed selection criteria
In both zones, the majority of farmers selected their seed for the next season from their harvest
(recycling). The criteria used for selection are summarized in Table 28.
A big cob was the most common selection criterion in both zones. However, size of the cob does not
necessarily reflect the genetic yield potential; cob size may be a result of the environment in which
the plant grew. For instance, a maize plant growing on border rows or on ant-hill soil may be larger
than others, even if it has lower yield potential. Therefore, farmers need to be educated about
effective seed selection practices.
4.9.2 Recycling improved maize varieties
Recycling the seed of improved varieties is a common practice in northern Tanzania. On average,
about 80% of sample farmers recycled improved varieties for 4 to 6 years. The type of varieties
planted between the two zones differed as expected. No sample farmer in the lowland zone reported
recycling the full season hybrid H622 or H632. Contrary to expectation, recycling of composites
was limited, probably because fewer farmers grew OPVs compared to imported and locally bred
hybrids. Only 18 farmers in the intermediate zone and two farmers in the lowland zone reported
recycling a composite variety (Table 29). The research recommendation is to recycle composites for a
maximum of three years. Hybrids are not recommended for recycling. The recycling procedure is to
select seed from the middle of the field when maize cobs have dried; those cobs have been exposed
to minimal cross-pollination from neighboring fields. Contrary to this recommendation, farmers
selected their seeds at home after harvesting the entire field.
In the intermediate zone alone, hybrids were recycled for a longer period than OPVs. On average,
hybrids were recycled for about 6.7 years compared to 3.9 years for OPVs. Imported varieties, such
as CG4141, were recycled for about 4 years in the intermediate zone and 2.5 years in the lowland
zone (Table 29).
Table 28. Seed selection criteria
Big cob Mature cob
Intermediate (N=79) 98.7 1.3
Lowland (N=13) 100.0 0.0
P-valuea 0.31 0.31
Table 29. Mean number of years of recycling improved
Variety Lowland Intermediate P value
Composites 8.5 (n=2) 3.9 (n=18) 0.15
H622/H632 0.0 (n=0) 6.7 (n=33) NA
CG4141 2.5 (n=8) 4.3 (n=21) 0.24
a Paired test comparing the two zones. a Paired test comparing the two zones.
Recycling is a result of high seed prices. Extension agents should discourage farmers from recycling
hybrids and advise farmers on how to recycle composites. Researchers should study the economics of
recycling improved varieties.
Farmers were also asked to state where they obtained their maize seed. About 77% said they
obtained seed from their own farms, while 11% got seeds from their neighbors. Twelve percent
bought seed from the market. This demonstrates that farmers are careful in obtaining seeds because
the majority of them obtained seed from their own fields to ensure that they knew what they were
5.0 CREDIT AND EXTENSION SERVICES
5.1 Credit availability
If small farmers are to be given a loan, such loans should be small to enable them to manage the
funds and repay the loan. Small loans to individual farmers, however, involve high administrative
costs that render them impractical. Commercial banks also require collateral that small farmers do
not have. Group loans are one alternative that has been suggested (Miller 1974). Banking
institutions in Tanzania have been giving loans to small farmers through their cooperatives (Due
1980). These loans have been used to buy inputs that are then sold to individual farmers. Farmers
growing export crops like coffee and tobacco have been given such inputs on loan. It was possible
to make these farmers repay their loans because they were obliged to sell their produce to the
same cooperatives. Until 1993 the cooperatives were the sole buyers of export crops. For food
farmers in Tanzania, there have not been any formal credit facilities other than NGOs. This is the
case in many developing countries. Miller (1975) observed that the major source of credit for small
maize and rice farmers in Nigeria was informal lenders like friends and relatives. Such loans were
small and their main use was for paying school fees and casual labor. Rarely were they used to buy
inputs. Anderson and Dillon (1992) note that these informal credit institutions play a vital role in
low income countries, but they are poorly documented.
Cooperatives often experienced liquidity problems that led to delays in input delivery and shortages
of inputs. Now export crop marketing has been liberalized, so farmers are not forced to sell their
produce to cooperative societies. This will probably make loan recovery difficult, resulting in a
termination of cooperative loans to small export crop farmers. The prevision of credit to small
farmers remains elusive, especially after the market reforms implemented by the government of
Tanzania. Commercial banks and other credit institutions have no plans for providing loans to
small farmers, especially small food crop farmers. Government intervention to make loans
available by forcing banks to waive collateral requirements (and other conditions) and make credit
available at below-market interest rates leads to inefficient and costly use of funds. Corruption has
also plagued such interventions; soft loans often wind up being more available to the rich than the
poor (Anderson and Dillon 1992).
SG-2000 has been the major source of credit for small food crop farmers in northern Tanzania.
The major drawback of credit from SG-2000 has been its short-term nature (three years) and weak
credit administration, which increased the number of loan defaulters when credit was extended to
a large number of farmers (Nkonya 1994). Nevertheless, SG-2000 is the foremost NGO
addressing the issue of credit for small food crop farmers in Tanzania. However, the program
would benefit by emulating the Grameen Bank in Bangladesh, which has been successful in giving
loans to small farmers on a sustainable basis. In recognition of the bank's success, its founder,
Yunus Mohammed received the 1994 World Food Prize (IPS 1994).
Sample farmers who received loans in the two zones represented 16% of the 126 respondents.
Table 30 shows that 19% of intermediate zone respondents received credit at least once compared
to 8% of respondents from the lowland zone. The difference in loan accessibility may be that the SG-
2000 project had more sites in the intermediate zone than in the lowland zone. The major source of
credit reported was NGOs, especially SG-2000. Only 1 out of 3 loan recipients in the lowland zone
reported receiving a loan from a formal bank. In the intermediate zone, 3 out of 17 loan recipients
obtained a loan from a formal bank.
Farmers did not report receiving any loans from informal money lenders, probably because they
thought that a 'loan' comes from a formal institution such as a bank or NGO. There was no
significant difference in the source of the loan between the two zones (p=0.05). Table 30 shows
some of the problems farmers encountered in trying to obtain loans.
Lack of collateral was the major problem farmers faced in trying to get loans from formal credit
institutions in both zones. Cumbersome loan procedures were cited as the second major problem.
Problems encountered in obtaining loans across the two zones were not statistically significant (p =
0.05). Table 30 should be interpreted with care because it involves a very small sample for the
5.2 Sources of information
There is a need to investigate farmers' sources of information about new technology in order to
assess the effectiveness of such sources in delivering technologies to farmers. Farmers were asked
whether they had received information on maize production technologies from various sources and if
they had adopted such technologies. Tables 31 and 32 summarize farmers' responses to these
questions. The three most important sources of production information cited were extension agents,
other farmers, and NGOs. In both zones, extension officers led in imparting knowledge to farmers
for all technologies, except draft animals in the lowland zone. The NGOs that have been active in
agricultural extension in northern Tanzania include: Sasakawa Global 2000 (SG2000), Farm Africa,
Heifer Project International (HPI), and religious organizations (e.g. ADDO), among others.
Table 30. Credit availability, credit sources, and problems of getting loans
Lowland Intermediate P value
Received credit? (% yes) 8.0 (n=36) 19.0 (n=90) 0.07
Source of credit: NGO 67.0 (n=3) 82.0 (n=17) >0.50
Bank 33.0 (n=3) 18.0 (n=17) >0.50
Problems of getting a loan:
No collateral 75.0 (n=32) 84.0 (n=74) 0.30
Cumbersome procedure 25.0 (n=32) 14.0 (n=74) 0.20
Other 0.0 (n=32) 3.0 (n=74) 0.13
a Paired test comparing the two zones.
The responses on sources of technology information were, however, biased against other farmers as
a source. A farmer may forget to report that he or she received production technology information
from another farmer because the two live together as peers, so the sharing of information sometimes
goes unnoticed. Notwithstanding this bias, 'other farmers' ranked as the second most important
source of production information for all technologies in the intermediate zone (Table 32). Use of
draft animals was learned at a tender age from parents, so it is not surprising that other farmers were
the most important source of information on use of draft animals in the lowland zones. Extension
efforts have been directed mainly toward improved varieties, chemical fertilizer use, and planting
configuration, which explains why over 80% of sample farmers in the intermediate zone and all
respondents in the lowland zone reported that they had received maize production information on
these technologies. Information on herbicide use had been received by only 7% of sample farmers,
implying that this technology was still new to the majority of farmers.
Technology on draft power was also poorly extended to farmers as only 14% of respondents in the
lowland zone and 41% of sample farmers in the intermediate zone had received information on this
technology. Draft animal technology is important because farmers experience labor bottlenecks
during planting and weeding. Using draft animal power could ease the bottlenecks and release labor
for other gainful activities. Ox-weeding technologies were practiced in some parts of Arumeru
districts for pure stand maize (Nkonya et al. 1991; Cunard, et al. 1983), however, this technology
was rarely practiced by farmers in other districts. Extension agents need to direct their efforts to
advising farmers on how to weed pure stand maize using oxen.
Table 31. Sources of maize production technology information for the lowland zone
Sources of information
Technology (% yes) Extension Other farmers NGO Other
Improved variety (n=18) 100 44 11 17 28
Planting method (n=21) 100 89 0 11 0
Fertilizer (n=16) 100 29 21 21 29
Weed management (n=14) 100 31 0 23 46
Pest/disease control (n=10) 100 56 33 0 11
Storage methods (n=14) 100 61 15 8 15
Animal draft (n=21) 14 0 67 0 33
Table 32. Sources of maize production technology information for the intermediate zone
Sources of information
Technology (% yes) Extension Other farmers NGO Other
Improved variety (n=72) 85 66 18 3 13
Planting method (n=69) 85 76 8 8 8
Fertilizer (n=65) 80 69 21 4 6
Weed management (n=64) 73 68 11 2 19
Pest/disease control (n=65) 74 79 11 4 6
Storage methods (n=61) 75 56 15 9 20
Animal draft (n=61) 41 48 24 24 4
Maize is the major staple crop in northern Tanzania. The NMRP has directed considerable efforts
toward generating technologies suitable for specified target zones. Northern Tanzania is categorized
by NMRP as an intermediate altitude zone, hence most of the varieties bred for the zone are OPVs
(Kilima, Staha, TMV1, UCA, and Tuxpefo). Two hybrids, H632 and H622, are also recommended
for the zone. This study observed that most farmers in the intermediate and lowland zones grew
imported hybrids, especially CG4141. The variety was preferred mostly because of its high yield and
early maturity, despite its poor pounding, storage, and taste characteristics. In the intermediate zone,
hybrids H632 and H622 were the second most preferred varieties followed by Kilima. In the
lowland zone, Kilima and Katumani were the second and third most preferred varieties, respectively.
This means the demand is lower for composites than for hybrids. However, the NMRP has released
more composites than hybrids for this zone. Maize seed production and marketing institution
inefficiency may be responsible for the low demand for composites, which are suitable for recycling.
More effort needs to be expended on breeding hybrids targeted for the Northern Zone. Another
strategy is to promote composite seed production at the village level to overcome seed marketing
The two-step Heckman analysis of factors affecting adoption showed that farming experience was
the only household characteristic variable that significantly influenced the probability to adopt
improved maize seed in the intermediate zone. The probable reason why other factors were not
detected was the high rate of adoption of improved maize seed. No factor influenced significantly
the intensity of adoption of improved maize seed. The probable reason for this may be the narrow
range of the level of adoption, i.e., the 1992-1994 average maize area under improved maize seed
ranged from 0.17 to 1.00 ha with a mean of 0.89 and a standard deviation of 0.21 ha.
All farmers in the two zones grew improved maize varieties in the 1994/95 season. The proportion
of area under improved varieties was 76% of total area under maize for both zones and methods of
land preparation in the same season. The average rate of adoption of improved maize seed between
1992 and 1994 was 94%. The rate of adoption increased dramatically between 1990 and 1994,
probably because of an improved input delivery system under liberalized markets and increased
extension efforts. However, about 80% of sample farmers recycled improved varieties, including
hybrids. In the intermediate zone, locally bred hybrids were recycled for a longer period than
composites. This is contrary to expectations and recommendations. Farmers in the lowland zone did
not usually grow the locally bred hybrids. Seed selection was performed at home instead of from the
middle of the field as recommended. There is a need to research the losses farmers incur when they
recycle hybrids and composites, as well as the economics of recycling improved varieties. The
extension services should also direct more effort to advising farmers on the best methods of
recycling composites and discourage them from recycling hybrids.
The rate of adoption of chemical fertilizer was still low. About 64% of intermediate zone and 44% of
lowland zone sample farmers applied chemical fertilizer. About 49% of intermediate zone
respondents used kraal manure in the 1994/95 season compared to 44% of sample farmers from
the lowland zone. The average rate of adoption of fertilizer application for the period 1992 to 1994
was 59%. The only farm characteristic that influenced adoption of fertilizer was livestock units. A
unit increase in livestock reduced the probability of adopting fertilizer by 1.5%. No household
characteristic significantly influenced the intensity of adoption of fertilizer.
Recommendations on fertilizer placement were poorly followed. The majority of farmers broadcast
or banded chemical fertilizer without covering it. More extension efforts should be directed toward
fertilizer technologies, as a majority of farmers use inefficient practices. Also, studies on the
economics of fertilizer use should be undertaken, especially now that input and output markets have
Crop residues are still poorly used by farmers in northern Tanzania. The majority of farmers fed
crop residues in situ, and a few in the intermediate zone cut and carried stover to feed zero-grazed
cows. This magnifies the importance of using chemical fertilizers. Rising fertilizer prices, however,
have forced farmers to use kraal manure and crop residues. Extension efforts advising farmers to use
organic manure to supplement chemical fertilizers should be increased. More research effort should
be directed to soil mining, supplementation of chemical fertilizers with different sources of organic
manure, crop residue management, and soil conservation. Additional fertility research will be
particulalry relevant because use of chemical fertilizer is likely to remain low in the foreseeable future
due to its increasing price.
Adoption of other recommendations, namely land preparation methods, frequency and time of
weeding and fertilizer application, and crop spacing have been readily and successfully adopted in
both zones. Hence, research and extension in those areas should be given low priority.
Formal credit is not available to maize farmers. With rising input prices, credit to farmers becomes
increasingly important. Therefore, policy makers and bankers should direct more effort toward
providing loans to small maize farmers in ways that will ensure a high rate of loan recovery and low
cost of credit administration. The experience of the Grameen bank of Bangladesh should be
Anandajayasekeram, P., A. Stroud and Z. Semgalawe. 1992. A Zonal Research Strategy and Work Programme for
Farming Systems Research -Northern Zone, Tanzania (Unpublished).
Akonaay, H.B., R. Tuni, T. Mmbaga, L.R. Chalamila and P. Xavery. 1994. Preliminary Information on Impact
Assessment on Maize Research/Extension in Tanzania (1974-1994) With Special Reference to the Northern
Anderson, J.R. and J.L. Dillon. 1992. Risk Analysis in Dryland Farming Systems. United Nations Food and
Agriculture Organization (FAO) Rome, pp. 9-38.
Anonymous. 1988. Maize Production Handbook. Tanzania Agricultural Research Organization (TARO), Dar es
Salaam Tanzania: 1 20.
Banfield. 1996. R. Personal Communication, Arusha.
Cunard, A.C., C.Y. Tang, P. Nyassi, L. Mushi (Mrs.), S.M.S. Masomo, J.A. Mamkwe, I.E. Swai and S. Nahum. 1985.
Interim Report. Results of the Diagnostic Survey of Arumeru District. TARO/USAID. Tanzania
Due, J.M. 1980. Agricultural Lending in Two African Nations. Illinois Research. Winter 1980 22(1):3-4.
Due, J.M. 1986. "Summary of Farming Systems Bean Research in Tanzania 1982-85." In: Minjas, A.N. and M.P.
Salema (eds). Bean Research Proceedings of the 4th Bean Research Workshop held at Sokoine University of
Agriculture. Interpress of Tanzania Ltd: 146 151.
Dunford, C 1980. Proposed Land Planning Units for Arusha Region, Tanzania. Arusha Planning and Village
Development Project & Regional Development Directorate. Earthscan Ltd, Nairobi.
FSD (Food Security Department). 1996. Tanzania Food Security Bulletin No. 2, April/May 1996. Ministry of
Agriculture, Dar es Salaam: pp. 6-7.
FSD. 1992. Comprehensive Food Security. Government of the United Republic of Tanzania and The United
Nations Food and Agriculture Organization (FAO) Ministry of Agriculture, Dar es Salaam: pp. 15-31.
Feder, G. R., R.E. Just and D. Zilberman. 1985. "Adoption of Agricultural Innovations in Developing Countries: A
Survey." Econ. Development and Cultural Change. 33:255-296.
Goodbody, S.1990. Maize, Time of Planting. A Review of Recent Work in Tanzania. In Moshi, A.J. and J.K.
Ransom (eds.) 1990. Maize Research in Tanzania. Proceedings of the First Tanzania National Maize Workshop,
Arusha, TARO, Dar-Es-Salaam.
Greene, W.H. 1993. Econometric Analysis. Second Edition. Macmillan Publishing Company, New York: pp 143
Haule, K.I.. 1990. "Soil Fertility and Fertilizer Use Research in Tanzania." In Moshi, A. J. and J. K. Ransom (eds).
Maize Research in Tanzania. Proceedings of the First Tanzania National Maize Research Workshop, Arusha,
TARO, Dar es Salaam: 127 140.
IPS. 1994. Inter Press Service News Release: Poverty Bank Wins. October, 12 1994. Washington D.C.
Kaliba, A., A. Featherstone and D. Norman. 1998. "A Stall-Feeding Management for Improved Cattle in Semi-Arid
Central Tanzania: Factors Influencing Adoption." Agricultural Economics (forthcoming in January 1998).
Kapaliswa, L., R. Gillis, E. Nkonya and P. Xavery. 1993. "The Economic Efficiency of Cereal Production for Food or
Beer in the Smallholder Sector of Northern Tanzania." In: Mwangi, W., D. Rohrbach and P. Heisey (eds.).
Cereal Grain Policy Analysis in the National Agricultural Research Systems of Eastern and Southern
Africa. CIMMYT SADC/ICRISAT. Addis Ababa, Ethiopia: 274-277.
Maddala, G.S. 1983. Limited-Dependent and Qualitative Variables in Econometrics. Econometric Society
Monograph No. 3. Cambridge: Cambridge University Press. 151-160.
Maliyamkono, T.L. and M.S. Bagachwa. 1990. The Second Economy in Tanzania. ESAURP, Dar es Salaam: pp. 2
Mdadila, J.M. 1995. Industry Review of Maize, Rice and Wheat, 1993/94. Marketing Development Bureau (MDB),
Ministry of Agriculture, Tanzania: pp. 23-32.
MDB (Marketing Development Bureau). 1993. Basic Data: Agricultural and Livestock Sector, 1986/87 -1991/92.
Statistical Unit, Planning and Marketing Division, Dar es Salaam: pp. 35-38, 40-43.
Matowo P. R., G. G. Mgema (1990a)" Optimum weeding in Maize in the Low and Intermediate areas of Tanzania."
In Moshi, A. J. and J. K. Ransom (eds) 1990. Maize Research in Tanzania. Proceedings of the First Tanzania
National Maize Research Workshop, Arusha, TARO, Dar-es-salaam, Tanzania: pp. 195-197.
Matowo P. R., G. G. Mgema (1990b) "Optimum Planting Densities and Fertilization for New Varieties of Maize in
Tanzania." In Moshi, A. J. and J. K. Ransom (eds) 1990. Maize Research in Tanzania. Proceedings of the
First Tanzania National Maize Research Workshop, Arusha, TARO, Dar es Salaam: pp. 198-204.
Mbiha, E.R. 1993. "Effects of Government Intervention in Maize Production and Marketing: Implications for
Agricultural Marketing and Price Liberalization in Tanzania." Agricultural Economics Analysis and Rural
Development vol 3(2).
Mduruma, Z.O., A.J. Moshi and P.R. Matowo. 1990c. "Use of Carbofuran as a Seed Dressing Insecticide for
Reducing Incidence of Streak Virus Disease in Maize (Zea mays L.)." In: Moshi, A. J. and J. K. Ransom (eds)
1990. Maize Research in Tanzania. Proceedings of the First Tanzania National Maize Research Workshop,
Arusha, TARO, Dar es Salaam, Tanzania: pp. 216-219.
Miller, L.F. 1974. Using Group Loans to Extend Production Credit to Small Farmers in Nigeria. Technical Report
AEE/74, Department of Agricultural Economics and Extension, University of Ibadan: pp. 7-20.
Miller, L.F. 1975. Present and Potential Use of Credit by Small Maize and Rice Farmers in Western and Kwara
States, Nigeria. Technical Report AEE/74.3, Department of Agricultural Economics and Extension, University
of Ibadan: pp. 43-54.
Missiaen, M., and K. Lindert. 1993. Tanzania: Agricultural policy reform. In: Rosen, S. 1993. Agricultural Policy
Reform. Issues and Implications for Africa. USDA Economic Research Service, Foreign Agricultural
Economic Report No. 250: 97-106.
Moshi, A.J., Z.O. Mduruma, N.G. Lyimo, W.F. Marandu and H.B. Akonaay. 1990d. "Maize Breeding for Target
Environments in Tanzania." In: Moshi, A.J. and J.K. Ransom, Eds. 1990. Maize Research in Tanzania:
Proceedings of the First Tanzania National Maize Research Workshop, Arusha, TARO, Dar es Salaam,
Tanzania: pp. 11-16.
Mowo, J., J. Floor, F. Kaihura, and J. Magogo (Eds). 1993. Review of Fertilizer Recommendations for Tanzania.
Part 2, Revised Fertilizer Recommendations for Tanzania. National Soil Services, Mlingano Tanga Tanzania.
Soil Fertility report No. F6: pp. 25-29.
Nkonya, E.M. 1994. Adoption of the Sasakawa Global 2000 Maize Production Package in the Marginal Areas of
Northern Tanzania. A thesis submitted in partial fulfillment of the requirement for the degree of M.S.
Department of Agricultural Economics, College of Agriculture, Kansas State University, Manhattan Kansas,
Nkonya, E.M., S.D. Lyimo, P. Sulumo, M.Z. Owenya, G. Modestus, P. Xavery, P. Mushi and N. Lukumay. 1991.
Final Report of the Arumeru District Survey. Selian Agricultural Research Institute, Arusha Tanzania.
Nkonya, E.M., P. Sulumo, L.C. Mushi, G. Modestus, F. Mngube, Masele and Msuya.1992. Mbulu Rapid Rural
Appraisal: Potential and Constraints of Crop Production in Mbulu District. (Unpublished).
Nkonya, E.M. T.M.F. Samkyi, D.L. Kessy and V.C. Akulumuka. 1988. Coffee Survey Preliminary Report. Lyamungu
Coffee Research Station (Unpublished).
Nkonya, E.M., T. Schroeder and D. Norman. 1997. "Factors Affecting Adoption of Improved Maize Seed and
Fertilizer in Northern Tanzania." Journal of Agricultural Economics, 48(1):1-12.
Nyaki, A., S.D. Lyimo and L. Mawenya. 1991. Workshop on Environment and the Poor Soil and Water
Management for sustainable smallholder development. Field exercise and tour guide in Tanzania 5 June
Nyambo, B.T. and J.C. Kabissa. 1990. "The Status of Maize Stalk Borers in Tanzania: A Review with Emphasis on
Possible Future Research Proposals." In: Moshi, A. J. and J. K. Ransom (eds) 1990. Maize Research in
Tanzania. Proceedings of the First Tanzania National Maize Research Workshop, Arusha, TARO, Dar es
Samki, J.K. and J.F. Harrop. 1984. Fertilizer Recommendations in Tanzania on District by District Basis.
National Soil Services, Mlingano A.R.I. Ministry of Agriculture, Dar es Salaam: pp. 16-40.
Smale, M., R.E. Just, and H.D. Leathers 1994. "Land Allocation in HYV Adoption Models: An Investigation of
Alternative Explanations." American Journal of Agricultural Economics, 76: 535-546.
Uronu, B. 1990. "A Review on the Distribution and Control Strategy for the Larger Grain Borer Prostephanus
truncatus (Horn) in Northern Tanzania." In: Moshi, A. J. and J. K. Ransom (eds) 1990. Maize Research in
Tanzania. Proceedings of the First Tanzania National Maize Research Workshop, Arusha, TARO, Dar es
Salaam, Tanzania: pp. 229-232.