UNITED NATIONS 
DEVELOPMENT PROGRAMME
Emergencies Unit for Ethiopia

A Linear Model for Predicting Enset Plant Yield
 and
Assessment of Kocho Production in Ethiopia


By Dr. Robert Shank and Chernet Ertiro

Prepared for

Table of Contents

SUMMARY OF BACKGROUND INFORMATION
 
Definition of Terms

The Need for Assessing Enset Production

  • Enset production data of the past
  • Attempts to measure Enset yields
  • Production assessment methods and survey techniques
  • The Enset Cropping Areas METHODOLOGY OF THE ENSET PLANT YIELD REGRESSION STUDY

    RESULTS AND DISCUSSION OF THE PLANT YIELD REGRESSION STUDY

    Factors that could bias the regression relationship METHODOLOGY OF ASSESSMENT OF HOUSEHOLD AND NATIONAL KOCHO PRODUCTION

    RESULTS AND DISCUSSION OF HOUSEHOLD AND NATIONAL KOCHO PRODUCTION

    Factors influencing Enset production and hectarage in Ethiopia Future Developments in Enset utilization List of Tables

    Table 1. Crop assessment and food aid needs for Ethiopia for 1993 and 1994.
    Table 2. Annual national food production in Ethiopia for the years 1987-1993.
    Table 3. Hectarage and production of Enset products in the Southern Ethhiopia Peoples Administrative Region in 1994.
    Table 4. Kocho yield statistics for individually measured plants by age and region.
    Table 5. Effect of transplanting on Enset growth and Kocho yield.
    Table 6. Enset cropping data from sample farms of the Southern Ethiopia Peoples Administrative Region in march, 1995.
    Table 7. Nutritive value of some common foods consumed in the Enset growing areas.
    Table 8. Distribution of Kocho plant samples by weight categories.
    Table 9. Enset plants measurement and derived Kocho weight data used in formulating the regression model.
    Table 10. The correlation matrix for plant Kocho yield with plant size characters over the 60 sample cases.
    Table 11. Estimation of plant Kocho yields using a range of plant measurements substituted into the regression equation.
    Table 12. Plants harvested and plant measurements and plantation hectarage for 100 households sampled by wereda, zone and region.
    Table 13. Enset production per plant, per household and per wereda from the survey data.
    Table 14. Reported Enset hectarage, production and production/ hectare as compared to assessment study data.
     

    List of Figures

    Figure 1. Parts of the Enset plant.
    Figure 2. Geographic map of the Enset growing area.
    Figure 3. Wereda map of the Enset growing area.
    Figure 4. Agroecological map of the Enset growing area.

    List of Graphs

    Graph 1. Graph of predicted Kocho production for various plant measurements using the regression equation.
    Diagram of the Enset Plants
     
     

    Summary and Background Information

    Enset ventricosum is the edible species of the separate genus of the banana family, thus named ‘false banana’, but the Enset fruit is not edible. Variation within the species to altitude, soil and climate has allowed widespread cultivation in the mid- to highlands of western Arsi-Bale, the Southern Peoples Nations Natinalities Regional State (SPNNRS), and western Oromia including West Shewa, Jima, Ilubabor and Welega. Three Enset derived foods are popular but the most common is Kocho. The plant is cut before flowering, the pseudostem (see definition below) and leaf midribs are scraped, the pulp is fermented for 10-15 days and finally steam-baked flat-bread is prepared. As many as 7 million people consume the low-protein Enset products as staple or co-staple foods, sometimes solely with Vitamin A foods but commonly without the needed protein supplement. Leaf, fibre and plant parts are used for food wrappers, cattle feed, ropes and house construction materials.

    In December 1993, the FAO/WFP Crop and Food Supply Assessment document provided estimates of Enset and root crop production for the past seven years amounting to about 10% of the net cereal and pulse production. However, Enset/root crop production was not included in the 1994 FAO/WFP Crop and Food Needs Assessment Mission report. According to the report, although its (i.e. Enset and root crops) contribution to nutrition is considerable, a formula for its inclusion in the food balances needs to be determined. The Mission felt that some confusion existed over recent Enset yield reports of 4 tons per hectare (in cereal equivalent) compared to Ministry of Agriculture estimates of 5 tons per hectare. Moreover, in the SNNPRS, 1994 estimates show that 300,000 hectares of Enset is projected to yield almost 10 tons per hectare. Inclusion of Enset production from Oromiya Region (Oromia) and the national root crop production would have placed estimated 'Enset and root crop production' at more than 1/4 of the total cereal and pulse production of Ethiopia. This would have created, on paper, a food surplus situation in Ethiopia that could endanger the food security of numerous communities, which are in fact food deficit areas.

    Anyone who travels in the Enset growing area would conclude that Enset is a major part of the food production/consumption system of the area but quantification of its complex production mechanism has eluded statistical measurement. Several reports give average plant yields but an 'average plant' varies across agro-ecologies, cultural systems and even households. The number of plants utilized by a household is reluctantly and often incorrectly revealed by the farmer. And furthermore, quantification of land allocated to Enset is confounded by the multi-year nature of the crop, harvested throughout the year, and a complicated system of multiple transplantings at ever increasing spacings. This is further complicated by the findings reported herein and elsewhere that immature harvesting occurs during periods of food and a large percentage of plant seedlings succumb to drought and pests.

    For this study, the authors hoped to circumvent these problems by constructing a regression relationship between the dimensions of the actual plant size utilized and the yield. This plant yield relationship was then used to multiply by the average number of plants harvested from a random sample of 100 farmers per wereda and then multiplied by the number of farmers in the Enset growing kebeles (peasant associations) in the wereda. Although land area was assessed, ascertaining production per unit land was circumvented by basing yields on households and their utilization.

    The average number of plants harvested per household per year varied from below 12 plants in a few weredas to more than 100 in several. If 50-60 plants are needed to fulfill a major staple food requirement, then obviously those harvesting the lower number must be supplementing the Enset foods and those harvesting the higher number are either surplus producers or could be utilizing smaller plants. Wereda per plant yields averaged 19.7 to 84.6 kg with an overall average of 44.2 kg at 50% moisture. This compares to farmers opinion that average yields are 25 kg and researchers reports of 12 to 42 kg at unspecified moisture levels. Annual household production of Kocho varied from around 1 quintal in low producing areas to more than 90 quintals in 3 weredas. Average annual household production varied from 5-15 quintals in subsistence producing weredas to 35-70 quintals in surplus producing weredas.

    Assessed total annual Kocho production from 80 weredas and 7 special weredas amounted to 4,381,903 mt (metric tonnes) of which 3,704,698 mt is from the Southern Region and 667,205 mt is from the Oromia Region. This compares to 5,422,935 mt reported by the Ministries of Agriculture of which 5,008,515 mt was estimated to be produced in the Southern Region and 414,420 in the Oromia Region. The previous production figure accepted by the FAO-MoA for crop assessment in 1995 was 2,482,574 mt but was derived by deduction and did not include production in Oromia Region.

    Pulse crops, which are annually assessed, account for only about 10% of the grain production in Ethiopia. They are the basis of non-meat national dishes, are a vital protein supplement to the cereal diet but tend to be very low in productivity. Unlike pulses, Enset and root crops are simple basic starch crops, quite low in protein, not very adaptable to varying food dishes but very high in productivity, and may account for more than 10% of the cereal equivalent consumption. The implications of heavy dependence on these "poor nutrition" crops may have serious implications on the physical and mental health of the people of the "Enset Culture."

    Therefore, with the derivation of a formula for assessing Enset production, Enset and root crop production should go back into the food equation. Acceptance of the herein assessed production figure would put the country and certainly the Southern Region in a surplus production situation. While Enset products are not easily transported or stored, other grains produced in the Enset growing area are and should be considered in the national food balance. It may be that balance sheets should be regionalized so that surplus and deficit stocks can be balanced according to traditional foods consumed in that region.

    The Enset production assessment does not contradict the current food deficit in the areas currently recieving relief aid. In fact the data supports the concept that, even with Enset, the land shortage in some areas has resulted in immature harvesting and does not allow the most efficient production of this of multi-year crop.

    Subsequent assessments, whether on individual farms or in entire affected areas, should show whether relief aid enables the household to recover in terms of allowing Enset plants to grow larger thus more productive.

    It is hoped that this report will help stimulate the Food and Agricultural Organisation (FAO) and the Government of Ethiopia to focus attention on this crop and conduct any future assessments as needed.
     

    Definition of Terms
     
    Agro-ecology a defined group of ecological factors (climate, soil, altitude etc.) that contribute to the classification of distinct agricultural system. 
    Amicho the fleshy inner portion of the Enset corm which may be cooked and eaten separately, tasting similar to potato. 
    Bulla the small amount of water-insoluble starchy product that may be separated from Kocho during processing by squeezing and decanting the liquid. It is eaten as porridge.
    Clone or variety a distinct type or grouping of plants within a species separable from other types by some form of heritable trait, be it visual, chemical or other. New plants of clones are usually reproduced asexually, i.e. without utilizing flowering and seed production. 
    Corm  an enlarged fleshy structure at the base of the plant from which new shoots emerge following destruction of the dominant shoot. 
    Cereal equivalent (CE) conversion of the yield or weight value of a food to the proportion of standard 3,500 calories/kg equivalents of cereals. Enset and root crop foods, being high in water, are adjusted by the following factors to give CE (see Table 7). 
    Kocho .54               Cassava .44 
    Bulla .57                  Yam .18 
    White Potato .21     Taro .32 
    Sweet Patato .32 
    Jump  the package of Kocho or Bulla ready for transport and marketing usually wrapped with dried leaf sheaths. 
    Kocho the pulp of the Enset pseudostem derived by scraping the individual pieces and excluding the fiberous remains. Bulla may or may not be extracted and the Amicho may or may not be included. The raw mash is chopped and fermenting and it is this product at 50% moisture that is being assessed. When steam baked, the flat-bread is also called Kocho. 
    Leaf sheath the basal part of the leaf which is wrapped around other leaf sheaths to form a false plant stem 
    Metric tonne  ten quintals or 1,000 kilograms 
    Midrib  the central portion of a leaf which supports and carries nutrients/water to and from the leaf surface 
    Multi-year crop  a crop which takes several years to mature or flower as opposed to perennial crops which bear fruit or seeds each year. Annual refers to crops which are replanted and harvested annually. 
    Plantation an area designated for grouping of crop plants such as Enset 
    Pseudostem   the ‘tree trunk’ formed by the bases of the leaves or leaf sheaths adhering to one another in concentric fashion 
    Peasant association or farmer kebele  an administrative grouping of farmers in a geographic district usually delineated by some natural boundary such as a foot path, a ridge or stream. 
    Sucker or Seedling- a new plant shoot that develops at the base of the leaf, in the case of Enset, at the juncture of the pseudostem and the corm, more properly called a sucker since it is not developed from a seed. 
    Wereda an administrative group of farmer kebeles usually having a common ethnic background and bounded by common geographic features. 
     

    The Importance of Assessing/Reporting Enset Production

    Governments need to take responsible action to ensure the avaialability of food supplies at all times. To do this, policies that encourage stability and growth of food production must be formulated in order to insure a reliable supply of inputs and provide protection against hazards. This includes a focus on research and extension activities that address those constraints limiting production, as well as a focus on other factors that could lead to increased output. These factors interact in determining the food security and nutritional status leading to the improved health and welfare of the nation's people.

    In addition, an accurate account of the vital food supply of a country is important to the food security of any people but especially in areas which have experienced food shortages in the past. To ignore the extensive production of Enset and its contribution to the food balance of this area would endanger the credibility of the World Food Programme and the Commission for Disaster Prevention and Preparedness’ appeal to donors for food aid. However, to accept the data provided by the Ministry of Agriculture in 1994 would have jeopardized any request for the required food aid (see Table 1). It has become increasingly evident that definite action must be taken to clarify the data utilized for planning food aid.
     

    Table 1. Crop assessment and food aid needs for Ethiopia.
     
    Assessment (Metric Tonnes) 
     
    1993
    1994 
    Cereal and pulse production 
    7,151,000 
    7,203,300
    Enset and root crop
    670,000
     2,747,635 for production 
    SNNPRS + ?? from Oromia 
    Total food available 
    7,601,000 
    10,530,935 
    Structural food required 
    8,626,000
    8,835,000(1995) 
    Structural food deficit
     1,025,000
     - 
    Structural food surplus
     1,695,935** 
     Actual food aid requested 
    882,000 
    Actual food aid distributed 
     911,000*
    606,980 (1995) 
     *revised **tentative, pending acceptance of SNNPRS Enset and root crop data
     

    Assessment Production Data- Past and Present

    A reliable Enset production assessment is difficult because plants:

    a) are transplanted 1-3 times at varying and increasing spacings,
    b) are harvested after 4-7 years depending on rainfall and the altitude conditioned growth rate;
    c) are consumed across seasons and years depending on a family's food needs; and
    d) vary greatly (agroecologically, in terms of variety, and individually) in yield of Kocho which is "about" 50% water and is not packaged in standard units.

    However, this is not reason to abandon estimation and research on the production of Enset. It is obvious that a reasonably accurate method of sampling could be used to assess production from the many reported surveys already available.

    Evans (1993) states that Enset crop production is a necessary parameter that needs to be measured since it is not only used for food security assessment of the Enset dependent areas but also for formulating government policy and research/extension resource allocation to the crop. In the words of one 1994 Crop and Food Needs Assessment Mission team member:

    "Everyone agrees that measurement of the Enset plantation area is straightforward and we have to rely on the farmer to truthfully report the number of plants he harvested. But that is where the consensus ends. What mathematical figure to use for the yield per plant and per hectare, no one agrees."  The assessment of Enset production for the years 1987 to 1990 was made by the Central Statistical Authority - CSA (Table 2) using enumerators trained in measuring yields from 2x2 meter samples of grain fields. For Enset production they had to rely on farmers' and limited research reports of average yield per plant, age of plants harvested, as well as estimated hectarage and plant spacing to calculate an across-the-country, per hectare-per year harvest yield. Following assessment-by-estimate in the early years, incremental increases were added without knowledge of actual changes in hectarage or production for the years 1989-92.
     
    Table 2. Annual national food production in Ethiopia for the crop years 1987 to 1993.
     
     Production  
    000 mt
    1987 1988 1989 1990 1991 1992   
    est.
    1993   
    fcast
    Cereals/Pulses   
    less feed, seed export and postharvest loss
       

    5,703

       

    5,763

       

    5,693

       

    6,231

       

    6,000

       

    6,755

       

    6,322

    Enset and Root   
    Crops-Cereal Eq
    570 570 570 600 620 640 670
    Milk and Milk   
    Products-CE
    249 261 273 285 300 300 330
    Meat and Eggs-CE 185 190 195 200 200 200 210
    Total Production 6,707 6,784 6,731 7,316 7,120 7,895 7,532
     Source: 1993 FAO/WFP Crop and Food Aid Assessment Report. FAO. Addis Ababa.

    Beginning in 1991, the responsibility for Enset assessment was shifted to the Ministry of Agriculture (MoA), first at the national level and recently to the wereda and zonal level. The local MoA staff estimate the hectarage and production of each Peasant Association or farmer kebele from each wereda. Regional decentralization and expanding emphasis on agriculture has resulted in shifting of staff and hiring of new recruits as agricultural extension agents, some of which have had little training and experience in assessment.

    In the 1993 MoA report compiled by CSA, 2,831,000 metric tons of Enset (1,529,000 mt-CE) were reported harvested from 137,000 hectares, giving an average yield of 206 quintal per hectare (11.1 mt-CE). This is almost three times the national yield reported by the 1992 WFP/FAO crop assessment for Enset AND root crops. In the Southern Region ( Gedio, Hadiya, North Omo, Sidamo and Kembata), produced 2,417,000 tons on 102,000 hectares at 237 quintals per hectare. In contrast, in the Oromiya Region (Borena, Jima and Bale zones) produced 414,000 tons on 35,000 hectares at 117 quintals per hectare.

    Climatic, edaphic (soil), cultural or varietal differences for the more than double difference in yield levels between the two regions are not known. However, administrative emphasis on the crop may have entered into the assessment since in the Southern Nations Nationalities Peoples' Regional State Enset is reported to be the major crop, accounting for 65% of the 37 million quintals of crop production, whereas in Oromiya Region Enset production amounts to 9% of the 46 million quintals of produced. Nevertheless, at the time it was felt that these figures must be incorrect, so an incremental increase (about 4.7%) was again used to estimate the 1993 production.

    For the 1994 crop assessment, data provided by SNNPRS MoA (Table 3) reported 5,008,000 tons of Kocho and Bulla production (2,299,000 mt-CE) on 300,000 hectares giving 166 quintals per hectare (8.96 mt/ha-CE). Sidamo moved from fourth to first in production changing from 227,000 mt on 23,000 hectares (82 qunitals per hectare) to 1,026,000 mt on 43,000 ha (240 qunitals per hectare). Gurage zone moved from seventh to second changing from 53,000 mt on an unspecified area to 864,000 mt on 52,000 hectares (166 quintals per hectare). Gedio zone moved from first to fourth in production changing from 889,000 mt harvested from 8,900 hectares (1000 quintals per hectare) to 186,000 mt on 18,800 hectares (100 quintals per hectare).

    The 1994 SNNPRS reported root crop production (white and sweet potato, taro, yam and cassava) of 825,173 mt on 80,181 hectares (103 quintals per hectare) gives a cereal equivalent of 248,435 mt for a total Enset and root crop yield of 2,747,000 mt-CE from SNNPRS alone as compared to the forecasted estimate of 670,000 mt production in the whole country. Production data on Enset and root crops in Oromiya Region was not collected. It became evident that a realistic assessment of production needs to be made before entering into the picture for the national food availability report. Including these crop figures would have significantly raised the caloric balance 30-50 calories per day for all Ethiopia. It was concluded that accepting the Enset and root crop data may be more misleading than no data at all, so only cereal and pulse production data were used in the determination of food aid needs at the time of the 1994 assessment (Table 1).

    Some of the descrepancies noted in the SNNPRS Agricultural Bureau Enset estimates can be attributed to the zonal and wereda staff of the MoA in their inexperience at estimation of hectares and yields. However MoA staff justified the hectarages in that past villagization programmes had discouraged planting of the multi-year crop whereas now people were moving back to their holdings and expanding their normal planting for food security. In addition new reports of higher per plant production, lesser spacing and earlier harvest were proclaiming higher yields potentials. During this time even expatriots were joining the ranks of those propounding Enset as a 'wonder' or 'miracle' plant with perhaps the highest food production potential of any crop.
     

    Table 3. Hectarage and Production of Enset products in the Southern
    Ethiopia Peoples Administrative Region in 1994
     
    Zone
    Hectares
    Kocho  (mt)
    Bulla (mt)
    Fiber (Qtl)
    Gurage
    52,400
    786,000
    78,600
    366,800
    Hadiya
    18,013
    210,752
    25,218
    108,078
    N. Omo
    55,609
    706,234
    72,292
    389,263
    Kembata  
    Alaba  
    Tembaro
    11,689
    136,843
    19,883
    93,568
    Gedeo
    18,844
    186,048
    NA
    NA 
    Shekacho
    3,279
    19,346
    NA
    NA
    Kaficho
    22,273 
    133,638
    NA
    NA
    Sidamo
    43,334
    940,347
    86,668
    476,674
    Bench
    750
    6,000
    NA
    NA
    Maji
    150
    1,005
    NA
    NA
    S/Omo
    210
    2,100
    NA
    NA
    Yem
    5,020
    62,750
    2,008
    10,040
    Amare
    4,000
    32,000
    NA
    16,000
    Buggi
    385
    3,080
    NA
    1,150
    Derashe
    373
    2,984
    186
    1,119
    Konso 
    80
    480
    NA
    NA
    Min Tea &  Coffee
    70,584
    1,023,468
    70,584
    282,336
    TOTAL
    300,000
    4,653,076
    355,439
    1,745,028
    Cereal Eq(mt)
     
    2,296,600
    202,600
     
     Source: Southern Ethiopia Peoples Regional Administration Ministry of Agriculture.

    Attempts to measure Enset yield

    In published studies on Enset production in sample areas, attempts to assess the exact production level have been numerous. Teketel (1975) processed a large number of plants of each of three varieties and found the average yield to vary from 27 to 34 kg/plant. In a 1985 extension cirrular, Dereje proposed from his studies that a family of five using Enset as the major staple food would consume 60 plants (4 to 8 years old) per year. Furthermore, at 2 X 2.5 meter spacing and 25 kg of product per plant, he estimates production from harvesting six year old plants would be 83 quintals per hectare. At 1900 calories per kilogram of Kocho, this would amount to a somewhat high estimated consumption of 1560 calories per day of Enset for each family member.

    In a study of 60 households in six villages nearby Attat Hospital in Gurage zone (Pijls et.al. 1994), 46 plants per year were harvested per household of 6.1 persons averaging 34 kg of Kocho per plant. The household average of 0.16 hectare of Enset divided into the number and yield of plants was found to give 95 quintals per hectare per year assuming 6 year-old plants were harvested. Average consumption of 0.55 kg per person per day amounts to 852 calories per day which was reported to be 78% of the carbohydrate intake. The high yield per plant in this study was thought to be influenced by the presence of cattle and the use of manure. Although it was known that sale of Enset occurrs, which would further reduce caloric consumption, no measurement of sales was attempted.

    In extensive descriptive characterizations of Enset culture collected from various sources including their own, Kefale and Sandford (1991) found North Omo Enset cultivation to average 78 square meters per household member or .69% of cultivated land. From Wolayita Agricultural Development Unit (WADU. 1977/78) data, the reported 78 plants harvested and processed averaged 27kg of Kocho per plant. Spacing of 18 plants within a plot gave Kocho+Amicho yields of 18-42 kg/plant which can be translated into yields of 33-70 quintals per hectare.

    However Sandford (personal communication) feels that drought, diseases, animal pests and harvest by choice or due to food shortage are reducing the number of plants at each age, the number of plants reaching maturity, and therefore the per plant as well as plantation yield. All these factors, if true on a widespread basis, would tend to reduce assessed yields. Therefore use of any figure for an area wide average mature plant yield would upwardly bias Enset yield assessment.
     
    Production assessment methods and survey techniques

    In 1993 FAO project ETH/86/013 (Evans, 1993) attempted to formulate a questionnaire to determine design and survey techniques for future measurement of Enset production, consumption, sales and contribution to diet. In this pilot survey, five households in each of two peasant associations in four weredas in each of five zones (195 samples) were selected for data collection. Although field size of each crop including Enset was enumerated, summarized results were not reported "because fallow and pasture land was not measured and would have biased size and proportion of landholding". Furthermore, data was taken in local land units (timad) and packages (jump) which were not standard and varied even within a community. (NOTE: Central Statistical Authority apparently has the raw data but plans to summarize or release are undecided).

    In the 195 households, number and age of Enset plants was taken. Total number of plants per holder varied from 87 in Illababor to 1191 in Sidamo. In the Enset plantation across all regions, 30% of the plants were new transplants whereas only 15% (85 plants per landholder) were over 5 years old.

    When trying to measure production of Enset products, it was found that "local unit names for measuring Kocho are varied" as is "the weight of the bundle for a particular unit, sometimes from 1 to 15 kg within the same community." Even after purchasing, harvesting, fermenting and weighing the produce of 163 mature plants individually, yield of Kocho varied from 8.5 kg per plant from 6 year-old plants from Sidamo to 92.5 kg for 5 year-old plants from Illubabor (Table 4). For the 163 plants yield tested, the average production was 30.8 but the standard deviation of yield from plants of the same age across the 5 regions averaged 9.3 kg. Although other extensive data was taken such as storage and preservation, livestock and poultry ownership, and foods consumed, the use of the criteria `age of plants' and `region grown' was found to be statictically unreliable for quantification of Enset assessment.
     

    Table 4. Kocho yield statistics for individually measured plants by age and region.1
     
    Region
     Plants Sampled
     Plant Age
     Kocho yield
    Standard deviation(kg)
     Coefficient of variation(%) 
    West Shoa
    1
    8
    5
    3
    2
    3
    1
    3
    3
    2
    1
    6
    7
    8
    9
    10
    12
    13
    14
    15
    16
    17
    23
    24
    23
    34
    17
    34
    20
    18
    33
    35
    35
    -
    9.45
    8.32
    2.39
    .99
    6.56
    -
    2.05
    6.95
    7.07
    -
    -
    39
    35
    9
    5
    19
    -
    11
    21
    20
    -
    South Shoa 
    5
    10
    12
    5
    5
    1
    4
    5
    6
    7
    8
    10
    46
    56
    36
    31
    37
    12
    14.08
    15.81
    24.88
    6.02
    11.99
    -
    30
    28
    69
    19
    32
    -
    Ilubabor 
     
    2
    10
    6
    2
    5
    2
    3
    4
    5
    6
    31
    47
    47
    93
    29
    3.54
    23.66
    10.39
    10.61
    14.17
    11
    50
    22
    11
    49
    North Omo 
     
    4
    13
    12
    4
    2
    3
    4
    5
    6
    7
    11
    26
    29
    31
    34
    5.64
    11.28
    11.21
    11.34
    .64
    52
    43
    38
    36
    1
    Sidamo 
    16
    5
    6
    6
    4
    5
    6
    7
    11
    22
    9
    33
    5.26
    11.02
    1.82
    14.89
    49
    49
    21
    54
     
    163
     
    30.8
    9.33
    30.48
     

    The study concludes that in determining Enset production, instead of trying to ascertain the age, spacing, area, number of mature plants or questioning producers on the number of bundles (jump) produced in a particular area or region, the best result is obtained by simply measuring the plantation, asking the number of plants harvested in a given year and multiply this by an average expected plant yield for an area or region. Still the basic deficiencies of this assessment method would be:

    1) mainly the determination of a basic plant yield considering the large variation in individual plant yield due to varietal, agroecological, husbandry and cultural differences as well as household processing,
    2) differences in age of harvest among households and during the crop year due to the current food supply and the plantation space available and
    3) basing the assessment on production per hectare, whether the area is measured or estimated, still involves the complex factors of plant spacing and number of times transplanted which varies widely across households and administrative areas.
    Hiebsch (personal communication) used the data from several sources to calculate estimated yield per hectare of different transplanting systems. In one method, the yield of a mature harvested plant is divided by the area and time that it occupied that space. In another method, when the system is assumed to be in equalibrium, the total annual yield is divided by the area occupied by that system. From reported plant yields and theoretical plant spacings over time, he estimates possible dry weight yields of 32 to 69 quintals/hectare/year (44.8 to 86.6 quintals @ 50% moisture).

    Recently, Yeshi (1995) compiled yield estimates of Enset production by accepting regional figures on Enset hectarage but not production. Instead for the calculation of production, she used an average of the per plant yields reported by 6 different researchers studies. This average of 27 kg/plant was multiplied by an estimated 1/4 of the Enset plantation area harvested per year and a calculated 400 plants (2.5 X 2.5 meter spacing) occupying that space. Finally that estimated production of 108Qt per hectare was multiplied by the reported 287,334 hectares of Enset in the Southern Ethiopia Peoples Administrative Region.

    The assessed production was given to be 2,482,574 mt per year but did not include production from Oromiya Region (Oromiya). An additional 15,487 hectares has been reported by Oromiya Region excluding Borena which is estimated by MoA to have 11,334 hectares. So this additional 26,821 hectares, at the same yield level, would have added an additional 289,667 mt bringing the total Ethiopian Enset production to 2,772,241 mt or 1,646,018 mt cereal equivalents.

    Using this method of assessment, the accuracy of the three factors, average plant yield, mature plants per hectare and hectarage, would continually be questioned and need costly re-evaluation. For instance if a farmers decided to increase planting of Enset, hectarage would immediately go up raising assessed production. However in fact the actual increase in production would not happen until years later. On the other hand, if some factor caused farmers to increase their harvest of Enset, hectarage would be reduced, assessed production would be reduced when in fact production/consumption has increased.

    Assessment of Enset production, in order to increase accuracy and reflect reality, needs to be uncoupled from the usual methods of annual crop assessment i.e. production/hectare X hectares grown. Annual crop production is uncomplicated by the fact that a defined area is planted, harvested and the produce consumed, sold or held in reserve within the specified season. Enset, however being a multi-year crop with the intermediate stages being edible, allows for flexible harvest timing and consumption. For this reason assessment could be based alternatively on utilization.

    Household utilization multiplied by number of households per administrative area is a valid assessment of production. Recent population census should assure reasonably accurate data on rural households. The remaining factor of household utilization should be assessable by sampling within the defined administrative area. In fact Yeshi (1995) suggests a nested design of sampling to insure greater randomness of sampling. A nested design would mean a certain percentage of households within a percentage of PA's within a percentage of weredas. The number of plants and the product for each size plant processed by households over time would need to be readily assessed. The number and size of plants by household can be enumerated by interview and verified by the number of plants presently being readied for future use. The processed yield per plant should be a function of the size of plant being utilized by the household.

    Non-destructive measurement of plant characters contributing to yield would appear to be the most logical means of assessing Enset plant yield. Kefale and Sandford (1991) while trying to use plant characters to distinguish varieties, report a reasonable 22 and 23% coefficient of variation for pseudostem girth and height respectively within plants of the same age/variety with sample size being only 3 plants. WADU (1979/80) report Kocho yields of 48 plants (24 each treatment) after measurements of pseudostem girth and height (Table 5). In this test the differences in measurable plant size were proportionately reflected in the Kocho yield. Furthermore many researchers continue to use these measurements to indicate growth response to a treatment and to characterize clones/varieties. This indicates not only measurable response to environment but also a heritable relationship.
     

    Table 5. Effect of transplanting on growth and yield of Kocho.
     
    Once
    Twice 
    Transplanting treatment 
    Plant base Circumference (cm) 
    170 
    191
    Pseudostem Height (cm) 
    218
    236 
    Kocho Yield (kg/plant) 
    20.7
     25.9 
     
    In summary this method of assessment would be similar to that utilized to estimate lumber yield from standing trees or meat yield of live animals. Simple, multiple or curvilinear regression relationships for product yield have been developed using diameter and length of a sample of trees in a forest or chest circumference of live animals.
     

    Succession of cropping systems determined by land economy

    Before going directly into the assessment study, it would be informative to understand the food security and land economy that interact to produce cropping systems and nutritional vulnerability in the Enset growing areas.

    Westphal (1975) describes the agricultural systems of the Enset growing Southwest Ethiopia as

    1. Staple food cropping i.e. Enset as a primary food crop (Gedio, Sidamo, Gurage, Hadiya, and Kembata zones)
    2. Co-staple food cropping i.e. Enset and cereals as primary food crops (Wolayita, Gofa, Keffa, Amaro and Yem zones)
    3. Grain Cropping i.e. cereals as the staple food with supplementation especially in off-saeasons with Enset foods (Welega, Jima, West Shoa, Illubabor and parts of Kefecho zone)
     4. Shifting cultivation i.e. rotational utilization of cleared forest areas with a variety of crops including Enset (Bench, Shakicho and forested Kefecho zones)

    In the Enset growing areas, these cropping systems are adapted to the land availability and the food production economy. In the far western part, where forest and grassland is not so limited, cattle herders grow Enset as food insurance. Plants as old as 17 years are reported and occasionally plants mature, form seeds, die and dry up without being harvested. In these areas, plants grow naturally `in the wild' having started from seeds dispersed from the maturing ones. Shifting cultivation and livestock products provide self-sufficient food supplies and distance from markets discourage cash cropping. These areas are seldom needing supplementary or drought-caused food aid.

    Also there are areas where the forest soils support either naturally occuring or cultivated coffee trees. Coffee prices were up recently (though coming down rapidly) and the income per unit land cultivated is good. Chat, banana, pineapple and other fruits are also common, along with livestock so desired cereals can be purchased from cash sales. Once land is cleared and where topography allows cultivation, along with Enset, cereal and pulse cropping is popular - wheat, barley and horse beans at the higher altitudes and maize, sorghum and hariocot beans at the mid-to lower elevations. These crops also produce residual fodder that support livestock in a supplementary role as draft power. However, in land short areas, cultivation of Enset and annual crops is by hand hoeing. Cattle, where affordable in these land short areas, are often staked in the small grazing areas and are slaughtered at maturity for meat consumption or sale.

    Nevertheless as size of holdings in the Enset growing areas decrease, as much as possible, farmers maintain cereal production but switch from low-yielding pulse cropping to high yielding root crop culture. Sweet potato and/or white potato are the main substitutes with taro, cassava, and yam being common to a lesser degree, probably for food variety and food security. The white potato is adapted to loose soils and higher elevations but very susceptible to the devestating late blight disease. Sweet potato is better adapted to the mid-altitude clay soils but can often be limited by shortened rainy seasons or drought.

    Vegetables and fruits, although still not large cash cropping systems for self-sustaining farmers, are beginning to occupy ever larger farming areas. Onions, cabbage, carrots and beet root patches are increasingly common and are supported by the Orthodox tradition of animal product fasting. Avocado, mango, koke (peach), ghesho (hop) and zatu (guava) seedlings are in high demand and MoA tree nurseries are beginning to supply the farmers. In some areas growing of Eucalyptus poles has become a common cash crop.

    Crops requiring high labour, but which have high production or value are the economic factors that would allow the densly populated areas to survive on small landholdings by generating cash for purchase of needed additional energy foods. However, these crops too are dependent upon market development, climatic dangers, high seed costs and expensive chemicals to control diseases and pests. These are among the reasons households increasingly depend on Enset for food security. In summary, the succession components in the farming systems as related to rural poverty, decreasing land holdings and increasing dependence on cropping technology can be summarized as follows:

    Livestock < Coffee < Pulses < Cereals < Enset < Root Crops< Vegetables < Fruit

    It may appear fatalistic to forecast that if Ethiopia's rural population and poverty continues to grow with limited land availability, the trend will be in the direction of more high yielding crops, poorer balance in terms of nutritional quality and quantity and higher risk of dependence on these few crops.
     

    Cropping systems and current food security/nutritional situation in four Enset dependent areas

    When considering the problems and solutions to the problems of the Enset growing area, it is important to consider the agro-ecology of the entire crop/livestock system and its means of supporting the family food security. However soil, climate, elevation and other factors effect the production potential and the solution of constraints in each area. The following is a preview of some agro-ecologies growing Enset and information sharing between them could be helpful in finding common solutions. From interviews with representative families, it is clear that by utilizing Enset, large families can derive a large quantity of foor from a small proportion of their land holdings (Table 6).
     

    Table 6. Enset cropping data from sample farms of the Southern
    Ethiopia Peoples Administrative Region in March,1995.
     
     Town  
    Zone
    Farm  
    size ha
    Enset 
    area ha
    Plants/yr  
    consumed
    Plant Yield  
    Kg
    Family size Plant  
    spacing
    Plant  
    Daime cm
    other foods Income  
    generation
                    grow buy  
    Dilla  
    Gedio
    1.5
    .5
    25
    10-15
    4
    2.0
    190
    maize s.pot fruit Coffee  
    Kocho
    Bule  
    Gedio
     
    .08
    14
    40-60
    12
    1.25
    250
    cereals  
    lentil  
    meat
    Cereals  
    livestock
    H.Selam  
    Sidamo
    1.5
    .07
    100
    6-10
    8
    1.0
    50
    wheat  
    onion  
    milk
    wheat  
    onion
    Geneto D  
    Sidamo
    .5
    .04
    12
    50
    7
    .75
    175
    maize pulses  
    fruit onion  
    S. cane
    tobacco  
    coffee  
    Euca trees  
    chat
    Durame  
    Kembata
    1.0
    .16
    100
    70
    12
    2.0
    160
    maize maize  
    s.pot wheat  
    w.pot
    coffee  
    chat  
    euca trees
    Boditi  

     

    1.25
    .04
    24
    75
    9
    .8-1.0
    145
    maize  
    s.pot  
    h.bean
    teff  
    s.cane
    Wajakero  
    Wolayita
    .75
    .12
    20-25
    100
    24
    1-4
    100
    maize wheat  
    s.pot barley  
    w.pot pulse
    coffee  
    teff  
    maize
    D.Wogane  
    Wolayita
    1.0
    .05
    40-50
    65
    12
    1.5
    125
    wheat maize  
    s.pot s.pot  
    w.pot taro  
    milk
    wheat  
    horsebean
    Waka  

     

    1.25
    .38
    18 
    100
    7
    1.0
    225
    meat maize  
    F.bean
     
    Angacha  
    Kembata
    1.25
    .35
    24
    35
    8
    1.5
    180
    wheat  
    F.bean
    wheat  
    F.bean
    Average
    1.10
    .17
    39.2
    58
    10.3
    1.6
    160
       
     

    1. The Enset/coffee/maize culture of Sidamo-Dilla
     
    This area is an intensive coffee growing area as a result of development work by the Ministry of Tea and Coffee. Fertilizer and pesticides have been made available for coffee growing, especially in the latter case for control of the Coffee Berry Disease to the extent that there are no honeybees in the area. With the current price of coffee being good, farmers can buy cereals and pulses as needed and there are no signs of food insecurity in the majority of the population. However, having experienced past fluctuations in coffee prices and production, farmers have become accustomed to growing their own Enset as insurance of food supply.
     

    2. The Enset/maize culture of Jima-Mizan Teferi
     
    Adequate and reliable levels of rainfall and length of the Krempt season have been benificial for this area and it has become well-known for high maize yields and surplus production. However oxen used to cultivate the maize fields, have consistently been plagued by Trypanosomiasis. Also because of the high rainfall, weeds and crop diseases restrict the use of small grain crops under these conditions. Enset is utilized during the cropping season food gap, is the source of the high energy foods needed by the farmer and is without the risk of disease or crop loss.
     

    3. The Enset/root crop/maize culture of Wolayita-Sodo

    Due to land pressures, livestock production has decreased and the area devoted to low yielding pulse production is also diminishing. Not only is the population dependent on these high yielding crops but they are grown and eaten successively within the same year. Maize can be planted in February and in July for two harvests in July (June for fresh maize) and November whereas sweet potato can be planted in October and February for harvest in January and May. Therefore, from late February until the end of May, Enset (Kocho) is the only available high energy food usually eaten with a little `Ubasha cabbage' as a source Vitamin A. White potato is planted at the higher altitudes in April to be eaten after June because late blight disease becomes severe with the heavy rains. Taro is harvested in November and Cassava is ready to eat in September.
     

    4. The Enset/Livestock culture of Dawro-Waka
     
    When traversing the river from Wolayita, the Omo valley is dry and stony with pockets of maize/sorghum and cotton until returning to the Enset cultivating highlands of Gofa, about 30km before Waka, the wereda center. These highland hills are steeply sloping and not suitable for cultivation. Although the soils are deep and rainfall percolation is rapid, erosion would be serious. The growing season is short, supporting only early maturing barley varieties and cabbage. For the three months of July-September it is reported to be only a little lighter in the daytime fog than at night. Pastures are green throughout the year though growth is not luxurious. Meat and milk products supplement the heavy consumption of Enset with no cereals or pulses except those purchased for holidays or by the wealthy. Although pneumonia and TB are present here, people choose to live here and take their cattle to work in the surrounding lowland because of the presence of malaria and trypanosomiasis there.
     

    Food Nutritional Comparisons for the Enset/Root Crop systems and the interacting factors of nutritional status
     
    During the food shortage of early 1994, it was reported that starving children from the lowland maize/sorghum cropping areas would recover much faster than those from Enset dependent areas. This is thought to be related to the higher incidence of Kwashiorkor related malnutrition cases. Unlike marasmus, the energy intake deficiency which is corrected by willing intake of caloric foods, Kwashiorkor is first symptomized by loss of appetite and apathetic unresponsiveness. Accompanying these symptoms, loss of immune response is followed by death from illness, usually diarrhea or malaria, unless protein intake is forceably administered.
     
    Table 7 lists the caloric and protein content of foods eaten in the Enset growing areas. The cause for protein deficiency during the food shortage-high Enset consumption period of February to May can be seen from the very low protein content of Enset and the root crops. The greatest risk is to children under the age of five who are weaned from their mothers milk. This is why vigilant and targeted supplementary food programmes are needed for poor and/or food insecure families in the Enset dependent areas. Areas which have cereal based consumption will probably recover with cereal relief supplies, but Enset/root crop-dependent populations will show high child mortality unless supplementary food programmes are carefully administered.
     
     

    Table 7. Nutrient value of some common foods consumed in the Enset growing areas.
     
      Food Source
    % Moisture
    Value of 100gm
    edible portion
     
     
    Calories
    % Protein
    Cereals  
      Maize  
      Wheat  
      Teff
     
    12 
    13 
    11
     
    363 
    344 
    345
     
    10.0 
    11.5 
    8.5
    Pulses  
    Horicot bean 
    Horse bean 
     
    10 
    9
     
    339 
    342
     
    24.0 
    25.9
    Oil crops  
    Noug (Niger)
     
    6
     
    513
     
    17.0
    Fleshy plants  
      Enset  
      White potato Sweet potato  
      Yam Cassava  
      Taro
     
    50 
    80 
    70 
    73 
    60 
    70
     
    190 
    75 
    114 
    41 
    153 
    113
     
    1.5 
    2.0 
    1.5 
    1.0 
    0.7 
    2.0
    Vegetables  
      Beetroot  
      Carrot  
      Leaf cabbage
     
    87
    90
    90
     
    45 
    33 
    28
     
    1.8 
    1.0 
    2.0
    Fruits  
    Avocado 
    Banana 
    Papaya
     
    75 
    70 
    89
     
    165 
    116 
    39
     
    1.5 
    1.0 
    0.6
    Animal Products  
    Beef 
    Fish 
    Eggs 
           Milk (cow)
     
    66 
    78 
    74 
    88
     
    202 
    95 
    158 
    64
     
    19.0 
    18.0 
    13.0 
    3.3
     Source: World Food Programme manuel on food nutritional contents; 1991, WFP, Rome.
     
    In a study carried out by USAID in Kenya (Neumann et.al. 1993) on the effects of food intake on toddlers and schoolers, limited quantity and quality of diet was found to cause growth retardation, reduced resistance to illnesses and impared cognitive function. In rural African areas, where large amounts of energy requiring work exist, no resting energy adaptations were found and stunting without recovery was prevalent. While mild illness was found to be related to household sanitation, parental literacy and illness exposure; low levels of food intake, particularly protein and fat, predicated serious illness. Apart from duration of schooling, diet quality (protein composition) was the second most important determinative of cognative performance in school. Students with good diet quality were also found to show more spontaneous social and leadership behaviors than their peers.
     
     
    Enset production reliability and improvements needed

    Effects of drought and rainfall on productivity of the system

    Enset grows best at altitudes above 1600 meters above sea level, not because it cannot withstand heat, but because it needs adequate soil moisture. Irrigation studies in hot climates show that it stores large amounts of water but wilts severely during long periods without rain. The fact that transplanting is done in the dry season, usually just before the start of the Belg rain, and that growth resumes but at a reduced rate throughout the dry season, has earned the crop the misnamed reputation as being drought tolerant. The fact is that in the higher elevations, ample soil moisture reserves are being depleted by this deep rooted plant. On the other hand, at lower elevations and on tight clay soils where soil moisture stress occur, severe leaf pruning is practiced to conserve plant water use. The practice of manuring, natural mulching by leaf and stem residues, and the high inception of rainfall by plant leaves all contribute to rainfall capture, soil moisture conservation and reduction of run-off as compared to bare-earth farming. Enset is therefore a reliable crop during seasonal rainfall shortages but would succomb to prolonged droughts.
     
    The effect of the dry period on carbohydrate reserves and caloric content of the Kocho has not been studied. Especially in the lower altitude and dryer climates, it would be expected that plant reserves would be depleted when growth rates decline. Under these circumstances advantage may be taken by harvesting and storing the Kocho supply for the dry period rather than harvesting month-by-month.
     

     Soil fertility and Enset crop improvement prospects
     
    Because of the practice of manuring, household waste composting and Enset residue mulching, soil fertility in the Enset plantation is much higher than adjacent fields and pastures. However, with decreasing pasture and forage resources, declining cattle numbers and available manure, Enset growth and fertility requirements are becoming apparent.
     
    The SNNPRS-MoA Enset team has decided to undertake fertilizer, spacing and variety yield trials. These will be quite difficult to interpret considering the 5-7 year length of the cropping season and the variance in plant yield even before treatments are imposed. Plot size of 4 plants with 3 replications could have above 20% inherent variation.

    Based on the limited data of the Institute of Agriculture Research, 6 Enset nursery sites have been selected for increasing certain 'selected improved clones' for distribution to farmers. These clones are reputed to be less susceptable to bacterial wilt and better adapted.
     
    Other agronomic practices should be critically examined for contributions to yield. For instance, transplanting is practiced to conserve land area but is known to delay growth to maturity by 1-2 years. The size of the trade-off between time and space in the two cropping systems is not known.
     
    Transplanting loss occurs (individual farmers transplant 2-3 plants per space) and some selection is probably practiced for better appearing plants. Grouping of transplants is probably practiced to fully utilize the limited supply of manure. Intercropping with competitive species such as maize, sorghum and climbing beans or already established coffee, hops or tree fruits may allow more efficient land utilization.
     

    Constraints to production from diseases and pests
     
    Bacterial wilt (Xanthomonas musacearum) appears to be the only major disease of Enset. While the disease is present and `of concern' in most Enset growing areas, farmers are not alarmed and appear to be living with the disease though asking for `medicine'. Extensive experiments were conducted by Dereje (1984) to explore the mode of transmission. One hundred percent infection occurred from cutting leaves with contaminated knives or injecting the bacteria onto the cut surface. Thirty to 60% infection was achived by dipping transplants or watering the soil surface with bacterial solutions. But no infection occurred when transplants replaced dead infected plants (time between replacement not specified). Bacteria were found to survive on the surface of contaminated knives for up to 3 days under humid conditions and up to 4 days under dry conditions.
     
    MoA pathologists and extentionists are instructing farmers in removal and burying of diseased plants and sterilization of knives with fire but many farmers believe other means of spread occur and do not cooperate. Presently it appears that most farmers simply grow additional plants and do not concern themselves with diseased plants.
     
    Other means of disease spread are still possible and the real means of spread should be elucidated. It is known that bacterial infection of plants occurs by contamination through wounds, breaks in the epidermis or through the stomates (natural ventilation pores). So being able to manually transmit the disease does not mean the same as how the disease is transmitted under farmers conditions. The as yet unanswered questions about Bacterial Wilt of Enset are:
     
    1. In many areas apart from the Gurage zone, successive transplanting and knife prunning are not practiced. So what is the means of infection of Enset plants in these areas? (Up to 10% infection was observed in these cases). Possible sources of Xanthomonas spread to new plants are

    a. Splashing rain
    b. Birds, bees/wasps, flies(fruit)
    c. Wind blown leaves of infected plants
    d. Lizards, snakes, etc
    e. soil inhabiting organisms (mole rats, insects, nematodes)
    f. rainfall runoff to newly planted corms or plants
     
     2. Is the farmers' practice of tying down the leaves, allowing the diseased plant to dry and then burning a sanitary control practice? It does prevent contanimation of adjacent plants that could occur during the recommended practice of removal of diseased plants.
     
    3. Wilt occurs most commonly in 3-4 year old plants. If wounding is the means of pathogen entry, then why are small plants not similarly affected?
     
    4. What is the threshold of bacteria necessary to incite the disease i.e. could small amounts of inoculum from the above listed natural sources be sufficient to cause disease? One milliliter of concentrated inoculum will incite symptom development within 2 weeks, however what is the effect of sub-threshold levels of inoculum over the long term.
     

    Another threatening pest, which can be more devastating to an individual household is the mole rat. These animals, the size of large rats, tunnel and move from plant to plant eating the roots and corm, inciting variable damage including immediate decline and death of the plant. Cases were observed where serious infestation had caused 10-20 plants/week to be salvaged for what little pulp was possible to extract. Control consists of flooding the tunnel and killing the emerging animals. A large fish-hook type snare trap with a string trigger through the tunnel is being introduced by the MoA.

     
    Area of Enset cropping
     
    The area utilized for Enset cropping encompasses vast areas of the Southwest quarter of Ethiopia (Figures 2 & 3). Although Enset is utilized throughout the area, population density is lower in the western part and Enset cultivation is not as extensive.
     
     The mid-highlands are termed Woina Dega and the highlands are called Dega. The area of Enset cropping correlates well with these altitudinal delineations (Figure 4). The lowlands or Kolla of the Rift valley, the Omo-Gibe River, the intermountain area of North and South Omo and the Southwest Ethiopian perimeter are too dry for this consumptive water-using crop.
     
    Enset is adapted to deep fertile soils and ample rainfall areas throughout Ethiopia but the derived foods are utilized mainly from Addis Ababa south and west. Enset plants occur in Hararge and Gojam but the large leaves are used there, here in Addis Ababa, as well as in the southwest as wrappers for baking flat breads.

       

    METHODOLOGY OF THE ENSET PLANT YIELD REGRESSION STUDY
     
    Following the lead of the WADU Study (1980) which indicated a correlation between plant size measurements (pseudostem base circumference and pseudostem height) and Kocho yield, a regression study of these variables was undertaken for the following reasons:
     

    1. These two parameters could easily be measured in the field both now and for future assessments by development agents as a non-destructive quantification of Kocho yield per plant regardless of size of plant utilized.
     
    2. The predicted yield of varying size plants multiplied by the number of each size harvested would lead directly to Enset production per household .
     
    3. Utilization of Enset at the household level could be related to food security and reflect fluctuations in annual crop harvested by the household.
     
    4.The problems currently encountered in trying to establish production per hectare over time would be avoided. These problems include documenting number, age and spacing of plants at each of up to 4 transplantings, number/size and time of harvest, and finally average plant production over highly variable agroecologies.
     
    5. Variations in zonal, ethnic and household husbandry result in variations in plant growth rate and ultimately yield per unit time and space.
     
    6. Due to the 3-7 year growth cycle and the flexibility of harvest time, changes in household food economy and land holdings which may affect Enset plantation size and Kocho production would be measured as area expansion/ contraction per se / poor yields / high yields, respectively.
     
    Using the predictive equation with the following parameters
     
     Standard Error =                variance
                                sum of observations - mean

    or SEe =  of an Estimate  and assuming an 80% correlation between plant size and Kocho yield, then  further assuming a range in harvested plant circumference from 150 to 300 centimeters and a range on Kocho yield from 10 to 70 kg/plant then
     
    Now choosing a 90% confidence interval of predicting yield, the 't' value on a two tailed graph would be 1.645 and solving for the number of plants tested would give Consequently it was decided that data on size and yield of 70 plants would be collected to form the model with an attempt to sample an equal number of plants throughout the expected range of yields. In addition, it was recognized that Kocho production from a certain size plant could also be influenced by variety or clone of plants, agroecological factors and household processing practices. For these reasons plants were sampled throughout the Enset growing area in order to get a random sampling of plant Kocho yield.
     
    After weighing the Kocho (Bulla and Amicho were included if present), a sample was taken for moisture correction and the woman indicated the size of plant and height of pseudostem included in the Kocho. Circumference was measured about 20 cm above the corm and did not include pruned or partial leaf sheaths. Pseudostem height was measured from the corm juncture to the cut-off point indicated by the woman. Cases sampled which had excessive fiber were not included in the analysis. Also cases removed from the analysis were those in which the plant had already flowered because extension of the pseudostem during the flowering process resulted in skewed expectation of Kocho production and high residual deviations from regression. In the final analysis data from 60 plants were used in deriving the model.
     

    RESULTS AND DISCUSSION OF ENSET PLANT YIELD REGRESSION STUDY
     
    As shown below (Table 8), even after purposefully measuring a few small and a few large plants, the majority of plants sampled yielded in the 10 to 60 kg/plant range. The actual plant measurements and moisture-
     

    Table 8. Distribution of Kocho plant samples by weight categories.
      Number of plants in each of the sub-classes sampled
     
    Yield of Kocho (kg)
     <10
     10-20
     20-30
     30-40
     40-50
     50-60
     60-70
     >70 
    No of plants sampled
     5
     13 
    15 
    9
     7
     7
     1 
    10 
     

    adjusted yields of the sampled plants is given in Table 9. As a result of the diverse area sampled, the plant yield and size, both circumference and psuedostem height, varied widely. There should be no attempt to correlate size of plant with any place or agro-ecology since random households were simply asked whether they have Kocho under fermentation. Also no attempt should be made to correlate altitude with any plant character for the same reason. Some processing structures were contaminated with flies, maggots and ants.
     

    Table 9. Data on Enset plants measured and weight of Kocho derived used in formulating the regression model.
     
      
    Zone
    Wereda/ 
    Village
    Kocho 
    Weight 
    (kg)
    Plant 
    Circum 
    (cm)
    P'stem 
    Height 
    (cm)
    Altitude 
    (m)
     
    Comments
    *Gedeo
    Wanago
    116.02
    225
    180
    2500
    chopped
    Arsi
    Kofele
    114.72
    298
    303
    2320
     
    "
    "
    112.59
    252
    270
    "
     
    *Gedeo
    Bule
    106.10
    225
    252
     
    flowered
    "
    "
    100.00
    237
    234
     
     
    N. Omo
    Chencha
    94.28
    346
    286
    2500
     
    "
    Bidessa
    92.65
    206
    214
    1980
     
    Gedeo
    Bule 
    91.60
    304
    300
     
     
    Jimma
    Goma
    87.99
    156.6
    177
     
    flowered
    Gurage
    Goro
    73.26
    157
    207
     
    chopped
    N. Omo
    Bidessa
    62.45
    218
    181
    1980
    moggots
    Jimma
    Saka
    57.68
    160
    243
    1750
    flowered
    W. Shewa
    B Girama
    53.47
    155
    234
     
    pair-6
    Keficho
    Chena
    53.38
    226
    187
    2120
     
    Sidama
    W Genet
    52.06
    189
    232
    1800
     
    Hadiya
    Konteb
    51.43
    232
    202
    2300
     
    N. Omo
    Boditi
    50.61
    150
    143
    1960
    pair-5
    W. Shewa
    B Girama
    50.28
    158
    208
     
    pair-6
    Jimma
    Deedoo
    49.80
    182
    162
     
     
    N. Omo
    Boditi
    49.39
    173
    196
    1960
    pair-5
    Keficho
    Chena
    43.99
    169
    165
    2120
     
    Shekecho
    Masha
    43.85
    171
    205
    2200
     
    Jimma
    Goma
    43.56
    132
    209
     
    flowered
    W. Shewa
    M Derra
    43.26
    209
    259
     
     
    Shekecho
    Yeki
    42.27
    180
    245
    1700
     
    Yem
    Sekora
    38.76
    167
    210
     
     
    W. Shewa
    Wonchi
    37.71
    214
    280
     
     
    Gurage
    Goro
    37.14
    157
    153
    1800
    pair-3
    W. Shewa
    M Derra
    36.14
    148
    230
     
     
    Keficho
    T Gesha
    35.11
    144
    127
    1860
     
    Sidama
    A/Tulu
    34.98
    104
    147
     
     
    Gurage
    Gumer
    33.67
    172
    176
    1900
     
    Shekecho
    Masha
    32.24
    195
    228
    2200
     
    Kefecho
    T Gesha
    31.67
    237
    152
    1860
     
    Gedeo
    Adame
    29.20
    142
    185
    2400
    pair-1
    Shekecho
    Masha
    29.23
    119
    170
    2200
     
    N. Omo
    Soddo
    26.94
    156
    187
    2280
    pair-4
    Gedeo
    Adame 
    26.72
    125
    210
    2400
    pair-1
    Hadiya
    Konteb
    26.45
    152
    175
    2340
     
    Jimma
    Deedoo
    25.65
    167
    183
     
     
    Sidama
    A Selam
    24.88
    218
    190
    2750
    pair-2
    "
    W Genet
    23.66
    163
    211
    1800
     
    Yem
    Sekora
    22.93
    207
    197
     
     
    N. Omo
    Soddo 
    21.84
    161
    166
    2280
    pair-4
    Gurage
    Goro
    21.63
    153
    174
    1800
    pair-3
    Keficho
    T.Gesha
    21.50
    204
    197
    1860
     
    Shekecho
    Masha
    21.50
    131
    183
    2200
     
    Sidama
    A/Tulu
    20.91
    124
    122
     
     
    N. Omo
    Soddo
    20.73
    160
    153
    2280
    pair-4
    Gurage
    Goro
    19.18
    137
    160
    1900
    maggots 
    W. Shewa
    Gura
    18.77
    171
    123
     
    flies/ants
    Jimma
    Gimbo
    18.77
    176
    252
    1700
     
    Sidama
    A Selam
    17.30
    142
    195
    2750
    pair-2
    "
    A/Tulu
    16.14
    140
    154
     
    pair-7
    "
    "
    15.76
    151
    153
     
    pair-7
    "
    "
    14.61
    142
    144
     
    pair-7
    Arsi
    Kofele
    13.97
    86
    120
    2320
     
    Sidama
    A/Tulu
    12.30
    73
    123
     
     
    "
    A/Wondo
    11.60
    99
    139
    1800
     
    Shekecho
    Masha
    10.75
    94
    157
    2200
     
    W. Shewa
    Gura
    10.61
    145
    177
     
     
    Sidama
    A/Tulu
    10.45
    74
    112
     
     
    "
    "
    8.84
    71
    139
     
     
    "
    "
    7.30
    107
    140
     
     
    "
    "
    6.07
    98
    109
     
     
    "
    "
    3.00
    58
    93
     
     
    W. Shewa
    Gura
    2.45
    136 
    100
     
     
     
    In Table 9 several plants are labeled as paired plants meaning that the sample plants were the same size, age and variety from neighboring plantations. These would be expected to yield equally except for household variations in husbandry and processing. Paired t-test analysis of these pairs of samples, as well as paired data from the calculated yield using the model, was conducted with the conclusion that the yields for paired plants from both sets of data were not significantly different. This indicates not only that same size/variety of plants yield similarly but also that processing, at least by neighboring households, is likely to result in similar Kocho yields from same age/size plants.
     
    In developing the model, the analysis of variance due to regression was highly significant. Testing for curvilinear and interaction (circumference and height) effects, these factors were not significant or useful in explaining yield response. However simple linear effects of both pseudostem circumference and height were highly significant in predicting Kocho yield. The correlations are given in Table 10 and the model is sufficient to explain 82% of the observed sample variations in Kocho yield.
     
    Table 10. The correlation matrix for plant Kocho yield with the plant characters
    pseudostem circumference and height over the 60 sample cases.
     
    variable 1. plant yield (kg)  
                 2. circumference (cm) 
                 3. height (cm) 
    1.0  
     .87  
    .83 
    var 1
    1.0  
     .77 
    var 2
     

    1.0 
    var 3

     

    Correlation of Kocho yield with pseudostem circumference was 87% and with pseudostem height was 68%. Correlation of circumference with height was 72%. The following equation thus describes the relationship of Kocho yield as a function of plant pseudostem circumference and height.
     
    Kocho Yield (kg/plant) = -32.1 + .26 X circum (cm) + .13 X height (cm)
     
     The response surface graph of this function is given in Graph 1 and a table for field conversion of the two variables to Kocho yield is given (Table 11). It should be noted that there is apparently some non-linearity in the real yield of the very small and the very big plants since the model predicts no yield of the very small plants which actually yielded 2-5 kg/plant. Also the model predicts less than the 100+kg/plant of the very big plants. This indicates additional sampling needed for those wishing to predict accurately these plant sizes in particular.
     

    Table 11. Estimation of Kocho yield per plant from various pseudostem circumferences and heights by application of the regression
    equation- Kocho yield (kg/plant) = -36.5 + .23 X Plant Base Circumference in cm + .19 X Pseudostem Height in cm.
     
      ht 
    cir
    90
    110
    130
    150
    170
    190
    210
    230
    250
    270
    290
    310
    60
    -
    -
    2.0
    5.8
    9.6
    -
     
     
     
     
     
     
    80
    -
    2.8
    6.6
    10.4
    14.2
    18.0
    -
    -
     
     
     
     
    100
    3.6
    7.4
    11.2
    15.0
    18.8
    22.6
    26.4
    30.2
    -
    -
     
     
    120
    8.2
    12.0
    15.8
    19.6
    23.4
    27.2
    31.0
    34.8
    38.6
    42.4
    -
    -
    140
    12.8
    16.6
    20.4
    24.2
    28.0
    31.8
    35.6
    39.4
    43.2
    47.0
    50.8
    54.6
    160
    17.4
    21.2
    25.0
    28.8
    32.6
    36.4
    40.2
    44.0
    47.8
    51.6
    55.4
    59.2
    180
    22.0
    25.8
    29.6
    33.4
    37.2
    41.0
    44.8
    48.6
    52.4
    56.2
    60.0
    63.8
    200
    26.6
    30.4
    34.2
    38.0
    41.8
    45.6
    49.4
    53.2
    57.0
    60.8
    64.6
    68.4
    220
    -
    35.0
    38.8
    42.6
    46.4
    50.2
    54.0
    57.8
    61.6
    65.4
    69.2
    73.0
    240
    -
    39.6
    43.4
    47.2
    51.0
    54.8
    58.6
    62.4
    66.2
    70.0
    73.8
    77.6
    260
     
    -
    48.0
    51.8
    55.6
    59.4
    63.2
    67.0
    70.8
    74.6
    78.4
    82.2
    280
     
     
    -
    56.4
    60.2
    64.0
    67.8
    71.6
    75.4
    79.2
    83.0
    86.8
    300
     
     
    -
    61.0
    64.8
    68.6
    72.4
    76.2
    80.0
    83.8
    87.6
    91.4
    320
     
     
     
    -
    69.4
    73.2
    77.0
    80.8
    84.6
    88.4
    92.2
    96.0
    340
     
     
     
     
    -
    77.8
    81.6
    85.4
    89.2
    93.0
    96.8
    100.6
     

    Graph 1

    Factors that could bias the regression relationship
     
     In an attempt to make the regression function applicable to the entire Enset growing area, samples were taken in as many weredas as possible and households were randomly selected within the wereda. The only qualifications for taking a sample were `Did they have Kocho of edible age in which all edible parts were intact?' i.e. had not been partially consumed. The use of this model was intended to be useful to predict, with relative accuracy, the per plant yield anywhere in the Enset growing area and under any cultural conditions.
     
    1. Bias caused by plant architecture
     
    There is variation in plant architecture affecting pseudostem measurements caused by climatic and cultural conditions interacting with expressions of the genetic variety. In warm climates with rainfall patterns and soil characteristics that result in low soil moisture availability, the pseudostem morphology tends to be short and squatty especially if plants are not crowded. This morphology is characteristic of the vertisol areas of Welisso and Shoa Region approaching the Awash basin. In moist humid areas with good soil moisture and longer growing seasons, plants tend to be more tall and slender especially if crowded. The tall slender plant type is commonly found in Sidama-Gedeo-Borena environs where plant placing is dense. Even here plants utilized vary from surplus seedlings of 80-100 cm circumference and 60-80 cm pseudostem height up to mature, well spaced and fertilized plants of 2.5-3 meter diameter and 3-4 meter pseudostem height.
     
    Altitude and the corresponding temperature/insolation conditions also affect plant architecture. High light intensity contributes to photosynthetic carbohydrate accumulation and, combined with good rainfall, causes lush growth and in general, greater pseudostem circumference. However the lower temperatures present delay maturity and draws out the age to harvest.

    Therefore it would be expected that the data from the samples collected in this study would apply with greater accuracy to the area in which those size samples were taken. For example the yield of very large plants for the model were largely determined at places of high altitude where large plants commonly occur. Conversely it was found that large numbers of small plants are often utilized in one area near Awassa and most of the small plant yield data was taken there. It is recommended that other researchers establish their own model if they need highly accurate data for a particular place of study. However considering the diversity of samples used in the present study, the authors consider the model sufficient to predict present and future yield assessments throughout the Enset growing area.
     
    It was pointed out by some Enset researchers that certain varieties tend to be more bulbous i.e. having greater circumference at or approaching the base. Certainly some varieties appear to have tall and straight or gently tapering pseudostems while others tend to show swelling at or above the junction of the corm. However since regression takes into account variations both in circumference and height, bulbous variations would not affect the model in the way one would expect from a model based on volume calculations.

    2. Variations caused from seasonal plant moisture content.

    The pseudostem of the Enset plant and more specifically the leaf sheaths that compose the pseudostem are composed of multitudinous, somewhat large mesophyll air/water spaces uniformly arranged between the upper and lower fiberous surfaces. It is not known to what extent the thickness of these spaces contract under soil moisture stress affecting the plant base circumference. While most of the Enset growing area does not experience drought, the normal period of absence of rainfall could affect soil moisture availability and plant water relationships. It is common practice in Gurage and adjacent areas to prune, even completely decapitate, plants during the 'dry season' to 'prevent dehydration' of plants but it is not known whether this is necessary or interrelated to the need for livestock forage. Certainly these conditions could affect the basal circumference of the plant.
     
    All measurements for this study, both the regression study and the survey, were made during the period from mid-September to the end of November, at which time plant moisture relations should have been stable. Normally crop assessments are made at this time of cereal crop maturing and so the model was designed to correlate with the timing of the need for this information. If yield assessments are needed for other times of the year, it would be necessary to check the effect of soil/plant moisture relationships on the plant circumference and therefore the equation.
     
    3. Variations caused by Kocho processing methodology
     
    Household processing undoubtedly affects kocho yields. Size of plant utilized would not bias model yield estimates since size is the basis of measurement but what is done or not done with any particular size plant varies tremendously among households. First the outer leaf sheaths which contain quantities of chlorophyll and anthocyanin pigments may or most likely would not be included in the Kocho. Secondly the inner leaf sheaths may be scraped hardily, including more fiber, or gingerly, giving better quality Kocho. Also the inner shoot apex which eventually becomes a flower primordia may or may not be included. Bulla may or may not be extracted and finally whether the corm is processed with the Kocho, wholly, partially or not at all. For the model, Kocho, Bulla and Amicho or corm were complete for each plant sampled and weighted as one. Obviously the moisture content of the Bulla and Amicho fractions, where present, differed from that of the Kocho subsample taken for adjustment. These fractions were present in about 20% of the 60 samples used for the model and it is hoped that by random sampling the processing variations were a random sample of the actual household processed yields.
     
    Out of concern for differences in plant yield being attributable to processing variations, several times where neighbors were sampled with similar sized plants, notation was made. Correlation between observed plant yields of these 'paired neighbors' was 87%. It can be extrapolated that variations in household culture and processing could contribute approximately 10% to variation in plant Kocho yield. Therefore, if particular ethnic groups are being assessed who utilize particular processing methodologies, their yield may differ slightly from the overall population.
     
     4. Variation caused by maturity of the plant.
     
    Since the model considers size and not maturity of the plant, changes in dry matter content at different stages of maturity could affect estimated plant yields. As the plant approaches maturity, larger amounts of starch are accumulated in the plant prior to flowering. It is commonly known that plants nearing maturity make the best quality Kocho and the highest yield of the highly prized Bulla. According to the model, large plants, whether approaching maturity or not would yield the same. It is not known how much difference in yield this would make except to say that yield of Bulla per plant is small, averaging 6% (Pjils et. al.).
     

    METHODOLOGY OF ASSESSMENT OF FARMER ENSET HECTARAGE AND KOCHO PRODUCTION IN ETHIOPIA
     
    During the process of Kocho sampling for construction of the model, farmer survey forms were distributed at the zonal and, as much as possible, directly to wereda level extension or crop production personnel (Figure 6). Seventy four weredas which were thought to be major Enset growing weredas were solicited to participate in the survey and meter tapes were provided to facilitate conducting the surveys. Instructions were given on how to approach the farmers and how to insure random sampling of 100 farmers per wereda. Obviously some weredas which were lesser producers in the major growing areas were omitted. Also some weredas were inadvertently omitted due to their being under the Ministry of Tea and Coffee. Segmo and Satema weredas were omitted because they were being transferred from Illibabor zone to Jimma zone. Other weredas were omitted either by the MoA or the authors due to their inaccessibility.
     
    Farmer reported number of plants harvested was expected to be biased in some cases so the number of plants available for harvest was also assessed by the enumerator. In cases where there was extreme discrepancy between these two, harvested number was adjusted. Measurement of mature plant pseudostem diameter and height was done on 2-3 plants per farmer. Farmer Enset plantation size was either asked, estimated, or actually measured. Since total household production rather than per hectare production was considered to be the aim of this survey, accurate measurement of the plantation was not considered to be critical.
     
    Plant measurements for each household were inserted in the model formula and multiplied by the number of plants harvested by that household. Average production as well as hectarage of the 100 surveyed households was multiplied by the total number of farm households in the wereda. In some cases, as noted in the tables, wereda production data was not recieved by the survey. For these cases adjacent weredas average production was multiplied by the number of farm households. Also in some weredas which do not grow Enset in greater than 90% of the households, only the number of farmers in Enset growing farmer kebeles or peasant associations was used to calculate wereda production.

    Figure 6. Sample survey form used to assess Enset production in 74 weredas
     
    Zone       Hadiya  Wereda               Soro 
    PA           1st Otoro  Farmers in PA     2401 
    Farmers in the Wereda        40,335 Hectares Enset in PA                       387 
    Hectares Enset in Wereda     4,281  Hectares Cropland in Wereda     63,426 
     
     

    3
    15
    64
    203
    250
    .17
     
    4
    35
    49
    187
    200
    .20
     
    5
    -
    20
    -
    .10
    immature
    6
    200 
    300
    190
    205
    1.80
    commercial
    7
    -
    -
    -
    -
    -
    no Enset
    8
    19
    200
    150
    185 
    .95
    expanding
    9
    40
    10
    125
    142
    .06
    poor farmer
    10
    37
    46
    228
    203
    .24
     
     

    RESULTS OF THE ENSET PRODUCTION ASSESSMENT SURVEY

    Of the 81 weredas in the Enset growing area of the Oromia and Southern Regions, 74 were asked to participate in the assessment survey. From these 74, 58 weredas responded by returning completed survey forms. Some important Enset producing weredas were missed because of the zonal MoA's failure to include the survey of weredas under the Ministry of Tea and Coffee Development. Other weredas were temporarily under transition from one zone/region to another.
     

    Assessment data on number/size of plants harvested

    The number and size of plants harvested by the household varied both among and within the weredas (Table 12). It is likely that harvest parameters reflect the household food supply and the harvest of annual crops. Number of plants harvested annually per household varied from below 12 in 4 weredas to more than 100 in 13 weredas. Also evident is the many weredas with large standard deviation of plants harvested, indicating there is considerable spread among households in the wereda in terms of above and below average number of plants harvested. Obviously, in the case of extensive harvest, above 50-60 plants per year, the family may not consume the entire product but are likely to be commercial producers as well, selling varying amounts through local, urban and Addis markets.
     
    Variation in the size of plant harvested is not as extensive as could be expected. Standard deviation of circumference and pseudostem height could have been high as this measure is expected to reflect not only age of plants harvested per se, but also varietal, cultural, and agroecological differences as well as household food economy differences. Field observations by the authors indicate that there are often a few large plants in reserve at some household for special purposes; it is possible that enumerators may have measured those few largest plants not yet harvested rather than the average plant harvested by the household.

    Several weredas which are known to be extensive Enset growers reported low numbers of plants harvested. This cannot be accounted for by accusing the farmers of minimizing their food production because the enumerator/development agent also counted and reported standing plants available for next year harvest. Rather the enumerator is likely to have minimized both the harvest reported by the farmer and the remaining mature plants. This is evident in 2 cases in Sidama zone (Awassa and Bensa), and in several cases of N. Omo zone (Chencha, Damot Gale, Kindo Koshia and Soddo) and Kefecho zone (Decha, Gawata, Gimbo and Tineshu Gesha). In these weredas, the reason why the number of plants harvested should be up to 5 times less than other weredas in the same zone with similar reported hectrarages is not logical.
     
    Again the size of plant harvested is reported to be dimensionally small in some weredas and therefore the resulting yield per plant and household production in these weredas is lower than would be expected. Two weredas in Sidama zone, Aleta Wondo and Arbegona, reported very small size plants being harvested even though harvest numbers were similar to that of the other weredas of the zone. Harvest of plants this size is not uncommon, especially for the corm food, Amicho, but not expected for the majority of the household Kocho consumption. It may be that enumerators should be instructed to differentiate between harvest for the corm food `Amicho' and that of the more mature pseudostem foods, Kocho and Bulla. In both the cases of unreasonably low numbers and size, the average of the other zonal weredas is proposed for the assessment (in parenthesis) until more extensive data can be obtained in these particular weredas.
     

    Assessment data on hectarage of Enset
     
    Although Enset hectarage was not used directly in calculating the assessment the data was taken to give indicative relationships. The small household plantation size reported emphasizes the productivity of Enset. It appears that 1/6th to 1/8th of a hectare is common among households. Even among what appears to be commercial producers, plantation size above 1/3 hectare is not common. In 5 weredas (Arodessa of Sidama, Yerga Chefe of Gedio, Telo of Kefecho, Agere Maryam of Borena and Welisso of West Shoa) unusually large areas were given for the Enset hectarage. While enumerators were not specifically instructed to measure the Enset area, they were asked to give estimates since crop area was used for assessment in the past. It is likely in these 5 cases, as may well be true for the other enumerators also, that estimating Enset area is more difficult and therefore less acurate than for cropland. It would be helpful in future assessments to have the hectarage reported in consistent units rather than as fractions, decimals or square meters, i.e. 1/4, .25 or 2500.
     

    General considerations in assessment data collection
     
    In general the three parameters needed for the assessment, i.e. plant numbers and harvest dimensions, were reported to be easily obtained by the enumerators-- usually development agents, MoA specialists or a combination of both. It is likely that specific enumerator instructions by survey experts would lead to better standardized and more accurate assessment. In this first assessment the authors directly contacted wereda experts, who were not necessarily the enumerators, in about half of the weredas.
     
    Verifying the availability of the 10-100 plants indicated by the household to be harvested annually took the enumerator a matter of a few minutes. Further, measurement of the pseudostem circumference and height of a few of those plants also required only a few minutes. The only materials needed was a 3 meter tape measure and the assessment could be accomplished by a single person without assistance and without harvest or destruction of any plants.
     
    While it was specified that the 100 households were to be a random sample, it was evident in a few cases that households with larger Enset production were enumerated first and the last few enumerated were notably smaller than the first. Another enumerator pattern that was evidebt was that of reporting plant numbers in multiples of 12, indicating that plants harvested per month was extrapolated for the year. This may have exagerated either the maximizing or the minimizing farmer and would not have included harvest for holidays and celebrations.
     

    Calculation of Kocho production per plant, per household and per wereda
     
    The regression equation for Kocho yield of varying size plants was included in the computerized data base to give production per plant (Table 13). The wereda average was weighted by the number of plants of each size. Apart from the two Sidama weredas, average plant yields varied from 19.7 kg in Bonke wereda of N. Omo zone to 84.6 in Yirga Chefe wereda of Gedio zone. Most commonly, assessed plant yields ranged from the high 20's to the mid 40's.
     
    The overall average of the wereda averages gave plant production of 44.2 kg. However this is not weighted by the extent of Enset plants harvested by the respective wereda. Those weredas where commercial interests allow plants to reach larger size are averaged equally with those weredas where large numbers of households consume plants of smaller size. Therefore it is misleading to use this figure for generalizing household production.
     
    Per household production was also calculated from the data base by multiplication of respective plant production for each household by the number of plants harvested for each of the 100 households sampled. Household Kocho production varied from a low of 78 kg in Seka wereda of Jimma zone to 9829 kg in Bule wereda and 9716 kg in Wanago wereda of Gedio. Several weredas had household production of 5-7000 kg but most commonly weredas using Kocho as a staple food on a subsistance level produced 600 to 1500 kg per household.
     
    Recent census data provided up to date figures on rural households which were used for calculation of wereda production from the household production. The average of the 100 households sampled in the wereda was multiplied by the number of rural households. In a few cases, where the wereda is on the edge of the Enset growing area, the proportion of the rural households using Enset was used instead of the total number. And in a few weredas that were not surveyed, the household production of adjacent weredas was used to estimate the wereda production (figures given in [ ] brackets).
     
    Assessed total annual Kocho production in the 80 weredas and 7 special weredas amounted to 4,381,903 mt (metric tonnes) of which 3,704,698 mt is from the Southern Region and 667,205 is from the Oromia Region (Table 14). This compares to 5,422,935 mt reported by the Regional Ministries of Agriculture of which 5,008,515 mt was estimated to be produced in the Southern Region and 414,420 in the Oromia Region. The previous production figure accepted by the FAO and National MoA for the crop assessment in 1995 was 2,482,574 mt but was derived by deduction and did not include production in Oromia Region.
     
    Since wereda populations are so large especially in the Enset growing area, small errors of the sample, in this case 100 households per wereda, are magnified thousands of times. Sample standard deviations could be used to calculate wereda variability in household Enset production and to project future sample size needed to assure desired survey accureacy.
     

    DISCUSSION OF THE ENSET PRODUCTION ASSESSMENT SURVEY
     
    Despite the great variability in plant age, size, and time of Enset harvest as well as household variability in culture and processing, the survey was successful in establishing a `formula' for assessing Enset contribution to the food balance in Ethiopia. Furthermore when conducted on an annual basis, as is the crops harvested annually, the Enset assessment proceedure utilized herein would reflect the household food situation in the wereda in relationship to other food crops as well as food shortages. From this aspect, this proposed assessment process would be useful to both the agriculturalist interested in production responses as well as the humanitarian interested in food security.
     
    Acceptance of the herein assessed Enset production figure would add considerably to the food balance sheet for Ethiopia. Enset/root crops, as well as livestock products, are deliberately left out of the equation because of the difficulty in getting a handle on production/consumption. However now that a `formula' for assessing Enset production has been devised, production of a crop that contributes 20% of the national carbohydrate and is the staple food of 18% of the population should be considered in the balance sheet. The country and certainly the Southern Region would be in a surplus production situation. Conversion of the assessed production to cereal equivalents gives 2,366,228 mt-CE which should be added to the 9,435,000 mt of cereal and pulses produced in 1995.
     
    It should be the philosophy of the crop assessment process to monitor the production season for the major crops in the major production areas. Other major food sources should not be ignored just for ease of calculation. Recent missions have avoided the contribution of Enset because it is not easily transferable, i.e. stored, transported and utilized by different cultures. The same could be said for non-users of teff, maize or wheat. The fact is that the other surplus carbohydrate from the Enset area, be it what may, could be `local purchased' and transported to the deficit area. This could support local agriculture similar to the local purchase of sorghum in the Humera area for the Tigre Region. It may be that the balance sheet could now be regionalized to see which food sources are in surplus and would be suitable to the food preferences of local or adjacent deficit area.
     
    The assessed Enset production does not contradict the occurrence of food deficits in Enset growing areas currently recieving relief assistance. In fact the methodology developed herein points to low household production of Enset in the Wolayita area as an indicator of food insecurity further agrevated by declining Enset use as a coping mechanism. Both the smaller plant size and the lower plant numbers harvested are indicators of the failure of Enset to fill the food needs. Whether the Enset plants are being harvested immature, the land area is too small for the household needs, deseases or pests are affecting production or poor soil fertility/ lack of manure is imparing growth needs ground verification. However the methodology developed should also be a useful tool for monitoring Enset production changes that could result from diverting family food consumption from Enset to food aid/employment generated income.
     
    In summary, in vulnerable areas which grow Enset, it is the food security stabilizing crop. Consumption can be adjusted to compensate for fluctuations of annual production. The methodology developed for Enset assessment, as it was intended, could be useful as an indicator, like food prices and terms of trade, both for increasing food insecurity and for recovery of the household. The historic annual record of size and number of Enset plants harvested would be an immediate as well as a trend indicator of vulnerability. The measure of recovery, however, is expected to be delayed due to the rate of Enset plant growth but would be a better indicator since it signals perseverence of the recovery.
      
    Household production parameters
     
    Plant harvest numbers, production per plant and per household reported in this study are consistent with isolated observations reported previously. As mentioned in the literature Dereje (1984) proposed 60 plants @ 25 kg per plant as the `model' Enset production for a household of six using Enset as a major staple food. Also Pjils et. al. (1994) reported family consumption (plus possible sale) to be 46 plants @ 34 kg/plant in Gurage zone, which is an area using Enset as a major staple food. The number of harvested plants recorded herein is similar to these reports for the staple food areas of Gurage and Hadiya zones. However, plants harvested is significantly lower in Kembata and N.Omo zones where food insecurity is known to exist. Plants harvested is also low in Keficho and Bench zones but reliance on other crops in these less densely populated zones probably explains the lack of food shortages. On the other hand, the much greater number of plants harvested in Sidama and Gedeo zones reflects a household surplus of Enset production. Consequently these zones are a known source of supply for the Addis and other urban markets.
     
    Extensive comparisons with plant yields reported in the literature are possible. While plant yields in this survey are uniformly reported at 50% moisture; many reports do not specify moisture content nor do they always include corm and Bulla portions in the Kocho. Yeshi reported the farmers concept of average plant yield to be 25 kg. However from reports of 12 to 42 kg, she uses 27 kg for the average plant yield for her assessment. Mekiso's report of an average 30.6 kg from 173 plants processed does not specify moisture content but is given as refined Kocho after a 25-73% post-fermentation weight loss. Of the 173 plants, 35 were from Sidamo and had `fermented product' yields of 73 kg but `refined Kocho' yields of 27 kg implying that the remainder was water. Evans and Pjils et al do not give moisture information relative to plant yield except the latter states that in a few Kocho samples, moisture ranged from 48-59%. The high plant yields reported here in some weredas is due entirely to plant size enumerated and not to water content of the products since correction for water content was included in the model.
     
    As the major staple food, per capita intake of Enset foods may be quite high even though the minimun energy requirements may still not be met. Pijls et al reported an Enset intake of .55 kg per person per day (1045 calories) which amounted to 82% of the total energy even though this is only 60% of the requirement (Which would be 2123 calories by their calculations). Earlier Selinus, as reported by Pijls, had reported Sidama region people consuming .80 kg per person per day which was 90% of total energy intake and would have supplied 1690 calories. At .55kg per person per day and per plant yield of 34 kg, they calculated a need for 36 plants to be harvested annually for the family of 6.1 (1225 kg/household/year). Therefore of the reported 46 plants harvested, they assume 10 were sold. Comparable figures for the reported Sidama consumption rate would have been 52.4 plants @ 34kg/plant or 1781 kg/household/year. The assessed production/consumption per household of subsistense level in this study is well below the Sidama consumption reported in 1971 and closer to the Gurage production/consumption reported in 1994.
     
    It is interesting to note the relationship between Enset hectarage and productivity. Pijls noted that the average hectarage in the Gurage study was .16 hectares and should be enlarged to about .27 to supply the household energy needs. Enset area reported in this survey is also in the .10-.20 hectare range for the weredas producing less than the 1500-1700 calorie per person (1800 kg/household/year) needed as a major staple food. Those weredas/zones producing above this level had Enset hectarages above .20.
     
    Immature harvest, food preferences and food shortages
     
    This suggests that there may be a critical mass for efficieny Enset production. Below .20 hectares per household using Enset as a major staple food, productivity erodes from immmature harvest. There may also be complex interactions with unknown effects in that the surplus producing weredas/zones also utilize the random spacing-mixed maturity method of culture rather than the highly organized arrangement of spacings and maturities. Those studying these factors i.e. Tiebebe and Sandford, (Kefale et.al. 1994) tend to feel that these interacting factors have brought the household to the the point of trying to meet short-term food needs without capability to consider long-term productivity. Land seldom lies uncropped in the Wolayita area which is already using multiple season cropping, relay cropping and intercropping even in the Enset.
     
    No data had previously been given on average plant size measurements for household Enset harvested or yield at different stages of immaturity. Considerable attention has been given recently to the high rate of attrition of plants under cultivation, with particular attention given to family food shortages. Early plant loss, i.e. 0-2 years of age, is often attributed to transplanting failure and disease/wild animal destruction. However extensive immature harvest is thought to indicate the family need for food before the plants have had time to reach the size for maximum yield. Plant size harvested and per plant yields appear to be greatest in the highest producing areas indicating the possible interaction of a number of factors: higher productivity, less food shortage and/or less disruption of the multi-year cropping through villagization, devillagization, migration, civil unrest etc.
     
    Farmers also reported immature harvest to provide a variety of foods (Kefale et.al. 1994) that is for the corm food. It is possible that a corm crop can be derived in the process of eliminating the slower growing transplants, i.e. during thinning. From this report it is also revealed there are certain clones which are cultivated especially for the taste of the corm rather than the productivity of the mature plant even in the food short Wolayita area. If this is true in a wide area of the Enset culture, a separate crop assessment may be needed to separate food shortage immature harvesting from food preferences.
     

    Irrelevance of the term Yield per Hectare in the case of Enset
     
    Because of the above mentioned factors, the complex patterns of Enset cropping and the intermittant use of this multi-year crop, it is of little meaning to know the yield per hectare. The fact that Enset could produce as large or larger quantities of energy food per hectare thatn other crops is established. But the more important fact is that with dwindling family land holdings, reliable production of energy food with flexible consumption potential is of great utility. Energy foods are essential to the daily existence of the family caretakers regardless of the current cropping environment. Annual crops deteorate in storage whereas Enset continues to grow until a few days before need necessitates harvest. No other crop provides the productive potential and the flexibility of consumption as that of Enset. Production per hectare at any one time in the year or across years will depend entirely upon the food status of the family relative to the production level of their other crops.
     
    This survey makes no allowance for urban production. Urban Enset growing for Kocho production is widespread even out of the Enset cropping areas such as Ginchi in West Shewa zone. Considering that with as small as 1/10th of a hectare can produce half of the family energy food is highly significant to urban centers with limited income generating potential. Ordinarily food production assessments do not need to include urban agriculture but Enset may be an exception.
     

    Contradictions between age structures and number of plant harvested
     
    Recent work on crop balance sheets for watersheds have raised new questions on the age structure of Enset plantations and the resulting effect on production. Sandford and Kefale (1994) report 41% plant loss of 1-2 year old plants and 13% loss of 2-3 year old plants. Only 41 plants over 3 years of age remained for consumption. The farmers reported drought and disease as the cause for this reduction. However, the authors suspected food shortage as the cause and conclude that lower wealth groups had lower percentage of the maturing aged plants. On the contrary when percent loss is calculated at each age using their data, even though the 'poor' farmers lost 71% of their 'seedling' aged plants, the 'very rich' also lost 48% of their 'seedlings' which at least in the latter case should not have related to food shortage and early harvest. We proposed that it is caused by drought, disease and land shortage for transplanting.
     
    Evans (1993) also reported a 60% reduction in plant numbers between the first and the fourth years of cultivation over the 5 zones assessed. This includes a 71% loss for Illubabor zone and a 64% loss for Sidamo zone neither of which would have been pressured to harvest early by food shortages. In her study, North Omo zone posted an 80% loss of plants between the first and fourth years.
     
    As mentioned again in the section on household financial security, in addition to the natural and transplant losses, intermediate size plants are harvested by the woman unknown to the man and independent of household food security status. Habte-Wolde et.al. as reported by Hiebsch (1995) reported that in 15 households in the Wolayita area, male and female farmers reported death as the primary cause of loss at first transplanting but harvest and utilization as the primary reason for reduction in plant numbers at the age of second transplanting and following. Hiebsch goes on to calculate a hypothetical Enset yield of 54 qt/ha/year, 13qt of which or 25% could be coming from harvest of immature plants. The authors concur on this point since harvest of intermediate-sized plants was often witnessed when interviewing the woman for the Regression study samples. However additional study would be required before the figure of 25% would seem in order.
     
    The additional 25% postulated by Hiebsch was calculated to come from a weighted average of an additional 400 plants harvested before the second transplanting and 120 additional plants harvested sometime before the final 46 plants matured yielding 34kg/plant using Pjils et.al. (1992) yield figures. Considering the number of small plants involved, even if the yield were low it should contribute more than 25% of the total yield.
     

    FACTORS INFLUENCING ENSET PRODUCTION AND HECTARAGE IN ETHIOPIA
     
    There are a number of factor influencing Enset popularity among farmers in Ethiopia. These factors are geared toward independent household survival in farming systems that have been fraught with pests, surprised by erratic rainfall and scourged by government villagization. Enset cultivation has been developed by independent farmer investigation and Enset foods have been utilized to provide consistent food availability to smooth over the rough times of household food security.
     
    It is likely that Enset culture developed as a result of droughts and insect plagues that destroyed the vegetation of the pastoralist in the lowlands. Bringing their cattle to the highlands, they needed crops providing energy foods that did not require annual culture and constant attention. Enset propagation suited them ideally in providing them a flexible food supply, a crop that competed well with present vegetation and was not seriously affected by droughts and pests.

    1. Enset food utilization and household food security
     
    Enset is basically an energy food that is low in protein and Vitamin A. As long as high protein animal products were available, the only shortage that would prevent a balanced diet was the vitamin A. This is probably why the `ubashaw' cabbage is so popular and highly utilized along with Enset foods. Many resource poor households, following exhaustion of their grain supplies, subsist largely by eating Enset and cabbage. Although the corm can be cooked and eaten like potatoes, it too is largely an energy food. The starchy Bulla, which can be extracted from the pulp by squeezing, is another low-protein Enset food that can be used as porage, giving some variety to the diet.
     
    However it is suspected that Enset is most popular because it is a `living refrigerator' from which the family can conveniently take as much food as it likes anytime during the season. During the process of sampling Enset plants it became evident that the family who has Enset and the woman who has Kocho under fermentation can easily remove a flexible amount of Kocho from the pit and steam it as required. If the men are working hard or unexpected guests arrive, it is a simple matter of going to the pit and taking out the required amount. None is wasted by having extra in the pit and more can be taken out if appetites request. Usually the female head of household can regulate when Kocho supplies are getting low enough to justify processing into a second pit to assure an uninterrupted food supply.
     
    When the authors stopped at a house near Ginchi, a traditional wheat and tef growing area in mid-October, the teff and wheat harvest was still months away. The only other food in the house was some maize nubbins recently harvested. The husband said they were partly of Gurage ancestry and since they usually experienced food shortage just when they needed energy for the harvest, they decided to grow Enset. The wife showed us her Kocho in a polypropolene bag in the house. The weather was too hot to store the Kocho in an outdoor pit but she readily made a patty that could have been baked on a fire in minutes.
     
    2. Household financial security provided by Enset
     
    Not only does Enset function in the picture of daily food supply and bridging the cropping seasons but also in household financial security against crop failures, to pay taxes and even to pay social obligations such as wedding expenses. A farmer likes to 'strut' his massive mature Enset plants near the compound as a sign that he is providing for his family and is a successful member of society. One farmer in Kofele wereda, just east of Sheshemene showed us 4 huge plants that he said he is keeping as a 'cash reserve'. His neighbor said that last year when he became sick he sold two plants, one to pay for his month in the hospital and the other for his wife to buy food. The arrangement explained to us was that he announces in the market that he has standing plants for sale at a certain price, after agreeing on a sale, the processing is done at his house with the wife assisting and guarding the Kocho pit. He concluded the discussion by saying that a 40 birr Enset plant was easier to replace than a 40 birr shoat.
     
    Another most important consideration in providing routine household financial security, is the frequent sale of small amounts of Enset by the housewife to buy family supplies at the market. Similarly to the woman's sale of grains by tins, but not regulated by the husband as is the case with grain, the woman can take small amounts of Kocho from the storage pit to sell for purchase of weekly supplies. One native described this as the woman's 'Sekret' in that the husband does not know the quantity of Enset utilized in this manner and she does not have to ask him. To this extent it is inferred that a major source of Kocho production, i.e. utilization of immmature plants, was discovered by the authors and should be assessed.
     
    When the woman, particularly in the non-defined spacing culture of Sidama-Gedeo-Wolayita, was asked to show the size of plant in the Kocho being weighed for the regression model, frequent revelation was made to the inclusion of small to medium sized plants in the Kocho. When the woman was asked why small plants were included, replies included:
     a) the plantation is too thick and needs thinning to allow room for others to grow bigger;
    b) Some plants were leaning over or damaged and needed to be used;
    c) Small plants are sweeter and more juicy which adds to the development of the fermentation process; and
    d) smaller plants are not so tough and thus are easier to process.
     
    But one of the major reasons for utilizing intermediate sized plants could be that the farmer is growing surplus numbers of this size and desires to maintain only those well spaced and growing more rapidly. In other words, he'll never miss some intermediate plants as long as plenty big ones are present. It is a silent agreement between the husband and wife that she can harvest as many intermediate size plants to boost Kocho production and pay for routine family expenses as long as she leaves enough big ones to give him a feeling of security. The implications for plant size/number and total household production of Kocho in discussed under the section on factors that could increase assessed production.
     
    3. De-villagization, Land Tenure and Land Resource Allocation
     
    Since the change of government in 1991, significant de-villagization has occurred resulting in establishment of new household plantations. Also land tenure issues while not giving the land to the farmers, assured him of not taking it away. This has created enough confidence on the part of the farmer to stimulate enlargement of the area devoted to the multi-year crop. Since accurate measurement of prior Enset hectarage is not known, it would be difficult to document hectarage changes.
     
    4. Free Market Systems and Commercial Production
     
    With governmental touting of a free-market economy, traders have developed Kocho collection and marketing systems. It was noted by the authors that divergent goals and production methods are developing among households and in certain geographic areas where commercial production is diverging from domestic production. It is commonly known that the highest amount of Kocho production per plant occurs just prior to flowering (though not necessarily the most per unit area). Current commercial production centers include Gedio zone and adjacent Borena weredas and Gurage zone with adjacent Shewa weredas of Welisso, Wenchi and Tikur Inchina. In these areas growers allow plants to reach a larger size and may include more plant material (outer leaf sheaths, longer pseudostems and more fiber) into the Kocho and are less likely to segregate Kocho qualities at the time of fermentation. If this practice is or becomes extensive, determination of a separate model with more extensive data from plants yielding 70 to 150 kg/plant should be formulated for commercial producers.
     
     A quick market survey in the Addis Ababa-Merkato was conducted by the authors. About 120 women reported selling 2 `Jump' of Bulla and 4 of Kocho each per week. They also said that people commonly buy an entire Jump of Kocho for a family holiday. This amounts to 25,000 Jumps of Kocho (2,000 mt) and 12,500 Jumps of Bulla per year. There are also smaller sattelite markets selling Enset products in Addia Ababa.
     
    Jumps of Kocho from Gedeo area were found to weigh an average of 79 kg (ranging from 60-100) and sell locally for 60-75 birr. Contract trucking from Welisso to Addis was reported to cost 10-15 birr per Jump. Bulla purchased wholesale for 4 birr/kg was sold for 6-7 birr and Kocho purchased for 2 birr/kg was sold for 4 birr. Kocho from anthocyanin pigmented plants was sold for 2-3 birr/kg and Kocho or Bulla with excessive fiber was sold for 2 birr less than the normal price. In conclusion, since Kocho can be purchased locally for less than 1 birr/kg and transported to Addis for 1-2 cents/kg, it should be economical to develop competitive markets for Kocho and its uses.
     
    5. Diseases and Pests of Enset
     
     The only disease currently threatening Enset crop production is Enset Bacterial Wilt caused by Xanthomonas musacearum. The wilt causes complete death of the plant within weeks of the first symptom, i.e. yellowing and drying of the emerging shoot. Individual plant infection and loss is common but 'hot spot' losses can amount to half of the harvestable plants. One farmer claimed that wilt consumes 3 times as much as his family. Pathologists and extensionists recommend uprooting and burrying of diseased plants as well as fire sterilization of knives, machetes and tools.
     
    Farmers believe that other means of spreading occur; in fact they have pointed out that in the process of removal from the plantation, brushing against other plants may contaminate them and spread the disease. Common practice of the farmer is to allow infected plants to dry in place and then burn them. It is likely that from the copious amounts of bacteria produced in infected plants, spread could occur naturally by splashing rain or passively by birds, insects and mole rats (gophers).
     
    Although the bacteria is highly pathogenic (a few drops of inoculum can infect and kill a large plant), at this point it does not appear to be highly infectious (spreading). This indicates that vectors or means of spreading are not extensive but does not assure that one will not develop. Care needs to be exercised in terms of vigilence for the development of vectors especially since the disease also affects banana. Screening of clones or varieties by researchers has not identified sources of resistance but rather differences in levels of susceptability have been noticed.

     

    FUTURE DEVELOPMENTS IN ENSET UTILIZATION AND INFLUENCE ON HECTARAGE
     
    1. Overpopulation, Rural Poverty and Cereal Price Parity
     
    Portions of the Enset growing area are among the densest populated rural areas in Ethiopia. Large family size has brought about farm subdivision until present land holdings in many zones amount to less than 1 hectare. While land holdings have decreased, proportion allocated to Enset and root crops production has increased. Even prior to subdivision and toukel construction, one can often see new plantings of the multi-year crop.
     
    Unlike other areas of Ethiopia which experience risks of rainfall failure and definitive growing seasons, soil moisture levels in the Enset growing area often allow for year-round annual cropping and even then, relay cropping is becoming common. The Enset growing area already is dependent on successive cropping of as many as three crops per year. Limited land resources must increasingly be devoted to highly productive, low protein quality Enset and root crops. Pulse crops, teff and even wheat are increasingly utilized for cash generation. Only maize and sorghum remain competitively priced with Enset and root crops. Families with low land resources are forced to grow and consume the most reliable and productive food crops. Enset is perceived to be among these crops and is often eaten 2-3 times per day throughout the year. A family with 50-60 mature Enset plants, which requires less than 1/10 hectare is thought to be food secure. It is no wonder that almost every household has a full complement of Enset plants at each stage of development.
     
    2. Spread of Enset growing to adjacent areas-historic records
     
    The Enset growing area is slowly enlarging into adjacent areas and has the potential for spread to other land short areas. Although the Gurage to Wolayita to Sidamo area has been considered the Enset growing area in the past, vast areas of Borena, West Shewa, Kefitcho-Sheketcho-Bench, and even Jima-Illubabor are now growing Enset. When the authors asked cereal growers of Jima and Shewa zones why they grow Enset, they replied that usually they had a shortage of food during the 'hungry gap' of June and July when grain stocks were exhausted and new crops were not yet mature. Now they can use Enset to fill in for whatever shortage of annual crops they encounter.
     
    Enset could be grown for food in many other parts of Ethiopia. Enset is found in scattered areas of East/West Hararge where it has been used for baking wrappers. Recently farmers in these areas have become interested in food consumption as well. Large areas of the mid-to highlands of northern Ethiopia could grow Enset also. Historical records of 16th century Portuguese priests accompanying the Orthodox in retreat from the Gran Mohammed show Enset cultivation as far north as Asmara (Dr. Steve Brandt, personal communication).

    3. Urban poverty, escalating food prices and traditions
     
    While rural overpopulation, land shortage and poverty encourage migration to urban areas, urban food prices are higher which forces the poor to seek the lowest priced foods. As grain prices rise, especially in periods of shortages, cheaper Enset foods could become available. Considering the productivity of Enset, increasing commercialization of growing, processing and marketing could decrease prices for Kocho below that of maize and sorghum. While non-fertilized maize and sorghum commonly yield 15-20 qt/ha, Enset yields of 40-70 qt/ha on an equal area-time basis are possible.
     
    Urban use of Enset foods is becomming increasingly popular. While Kocho used to be popular only for the Gurage and `southerner', it is now eaten, though to a more limited extent, by all ethnic groups. It is becoming traditional to serve Kitfo (raw meat) with Kocho at holidays, weddings and specialty restaurants. Some urban restaurants now serve wheat bread, injera and Kocho with meals. Merchants in the Merkato report individual customers buy an entire 'Jump' of 60-100 kg of Kocho for the Meskel holiday season.
     
    4. Enset in mixed flours for injera and other products
     
    Already in the Dembidolo area Enset is dried, ground and mixed with teff and other flours to make injera. Tests of Enset flour mixed with other cereals have produced consumer acceptable products with tef, maize, sorghum and wheat flours. Although commercial drying may add to the cost, Enset flours should be economical to develop.

    5. Soil fertility, cost of fertilizer and land degradation
     
    Because of the high rainfall and the slope of the fields in the mid- to highlands areas, continuous cultivation leads to depletion of soil organic matter, loss of rainfall percolation, erosion and land degradation. Even if farmers are careful to avoid erosion with conservation measures, the soil fertility has depleted over time and yields are low unless fertilizer is applied. Fertilizer costs have gone up substantially and are of questionable use for farmers and crops which are not market oriented. That is, unless product is brought to market to pay for fertilizer costs, use is not sustainable.
     
    Enset however is a soil building crop. Since Enset is a multi-year crop, tilling the soil is not necessary, the extensive plant canopy intercepts the rainfall and erosion is minimized. Even though manure is applied in some areas, Enset culture builds soil organic matter and soil fertility. Asnakech (personal communication) reports that soil structure, organic matter and fertility differ greatly in the Enset plantation as compared to adjacent fields or even pastures. In many cases plant residue from weeding and thinning/pruning of Enset are used as plantation mulch further raising organic matter and reducing erosion. In conclusion, Enset culture is more beneficial for long term sustainable, subsistence farming. Therefore land degradation and high fertilizer prices may favorably influence Enset production and hectarage.

    6. Land shortage and livestock feed
     
    As farmland is subdivided and cropland is needed, land available for livestock grazing becomes limited. Even common grazing pastures are being broken where suitable for cropland. As forage becomes limiting farmers are increasingly looking to crop residues for livestock feed. Enset as a crop produces high tonnage of forage along with the human food products. Already in most Enset growing areas, pruned leaves, thinned plants and harvested plant tops are utilized for livestock forage. Also in areas where dry season forage becomes limiting, successive pruning of Enset leaves to the point of complete defoliation of the plant is practiced to provide forage.
     
    Since the technology of producing livestock silage is similar to that of fermenting Kocho, additional forage could be stored from the growing seasons. Parts of the Enset growing area experience a rainy season with surplus vegetation followed by a dry season with a shortage of forage. In these areas Enset plant material not used for Kocho could be chopped into old fermentation pits and stored for dry season livestock utilization. Silage production does not increase nutritive value but simply stores and maintains its value for later use. In conclusion overpopulation and land shortages are forcing reduction of livestock numbers at a time when the beneficial effects in terms of Enset supplementing protein foods are most needed. Increasing the efficiency of seasonal surplus forage utilization by storing in earthen silos would enable greater production of the urgently needed livestock proteins.
     
    7. Intercropping Enset with legumes and cereals
     
    Although Enset is a very competitive and densely canopied crop, there are periods of time after transplanting and pruning when large amounts of space are under-utilized at the beginning of the growing season. Short season crops like pulses could be cultivated at this time, adding nitrogen to the soil, before Enset growth closes the canopy. Tall crops like maize and sorghum, which can become competitively established, are already intercropped with Enset in Wolayita-Hadiya-Kembata. Climbing annual and perennial legume crops are also a possible contribution to soil nitrogen and protein foods. Winged beans, climbing cowpeas and beans and local forage legumes could be adapted to benefit Enset plantation fertility without reducing but hopefully increasing Enset growth and production.
     

    References
     
    Dereje Ashagari. 1984. studies on the bacterial wilt of Enset and prospects for its control. Ethiopian Agricultural Research Conference. April. Addis Ababa.

    Evans, Valerie. 1993. Consultants report for Enset pilot survey. FAO. Addis Ababa.
     
    Hiebsch, Clifton. (personal communication) Professor of Agronomy. University of Florida.

    Kefale Alemu and Stephen Sandford. 1991. Enset in North Omo Region. Farm Africa. Addis Ababa.
     
    Kefale Alemu et. al. 1994. Gender, structure and landraces in peasant Enset Plantations in North Omo. Farm Africa. Addis Ababa.

    Neumann, Charlotte, et. al. 1993. Diet quantity and quality: Functional effects on rural Kenya families. Human nutrition CRSP. US/AID.

    Pijls, Loek, Arnold Timmer, Zewdie W/Gebriel and Clive West. 1994. Cultivation, preparation and consumption of Enset in Ethiopia. J.Sci.Food Agric. July 16.

    Shank, Robert. 1994. The Enset culture: Technical report on Enset or the `false banana'. Unithe Nations-Emergencies Unit for Ethiopia.

    Taye Bezuneh. 1983. Evaluation of some Enset clones for food yield with emphasis on the effect of length of fermentation on carbohydrate and calcium content.

    Tropical Agriculture 61:111.

    Teketel Makiso. 1975. A study on the yield of Enset. Proc. 6th IAR research seminar. Addis Ababa.

    WADU. 1979/80. Crops and pasture section agronomic report. Wolayita, Soddo.

    Westphal, E. 1975. Agricultural systems in Ethiopia. Ag Research Report 826.
     
    Yeshi Chiche. 1995. Assessment of Enset and root crops contribution to food supply in Ethiopia. FAO. Addis Ababa.
     

    Plate 1. Enset Cropping is usually near the house for convenience of processing by the women and manure appli- cation. Right: Enset between the house and the cereal fields near Ambo, West Shewa zone of Oromia. Center: Enset behind the house but after the garden and seedling nursery near Wendo, Sidama zone of SNNPRS. Bottom: Enset completely surrounding the house near Bule, Gedeo zone of SNNPRS.
     
    Plate 2. Enset seedlings come from suckers produced on corms. Top: First year seedlings are transplanted close together, while 2nd and 3rd year trans-plants are spaced increasingly wider (background) Gumer of Gurage zone, SNNPRS. Center:In other cultural methods which minimize transplanting, seed-lings may be periodically thinned and consumed, fed to livestock or used for mulch and weed control. Agere Selam, Sidama zone of SNNPRS. Bottom: Seedlings are grown commer-cially in highland areas of Sidama and Gedeo zones of SNNPRS.
     
    Plate 3. Processing is labour intensive for household and village women. Top: The pseudo-stem pieces are separated and halved longitudinally. Bottom right: Holding one piece with her foot, the inner surface is scraped with a half of a bamboo stick. Bottom left: The pulp is hand and/or foot packed into a leaf linedsurface of earthen pit for fermentation.
     
    Plate 4. Enset products can be stored and eaten as needed or bundled in dry leaf sheaths for transport to market. Top: A surface fermentation pit is uncovered showing a packet of starchy Bulla on top, more fresh leaves and then the Kocho - here darkened by the purple-red anthocyanin plant pigment evident in the plant variety behind the lady. Bottom: Packages of Kocho called `Jump' weighing 60-100kg are collected by a `middle man' awaiting transport to retail sale markets.
     
    Plate 5. Pests affecting Enset are few but can be devastating to household food supplies. Top: A woman stands in front of a bacterial wilt infected plant, first evident from the yellowed and drying new leaves. (Brown leaf tips and margins in the background are caused by frost.) Bottom left: Pseudostem leaf sheaths, normally clean and white, are filled with putrid bacterial slime plugging the vascular water flow to leaves. Bottom right: Dying plants may be uprooted leaving blank spaces in the plantation or they may be left to dry and burned in place.

    Plate 6. Top left: Mole rats or gophers tunnel to the succulent corms especially in the dry season. Top right: Damaged by mole rat feeding, corm rot may set in, seriously retarding corm and top growth. Bottom: In systems that utilize successive transplanting into the same hole, pest build-up results in serious plantation space under-utilization and food crop loss.
     
     
    Additional resource persons on Enset cultivation, production and utilization.
     
    General culture  
    Ato Chernet Ertiro 
    Former Enset Team Leader 
    Ato Girma 
    Horticulture Team Leader 
    SNNPRS-MoA 
    Awassa Ph 20-10-77 
     
    Ato Esato or Ato Berhanu 
     Soddo wereda MoA 
    Wolayita, Soddo 
    Phone # 67 
    Diseases and pests  
    Ato Gizachew 
    Inst of Ag Res 
    Ato Dereje Ashagari 
    Sholla Plant Protection 
    Awassa Ph 18-36-71 
    Mechanization  
    Awassa College 
    Inst of Ag Res, Melkassa 
    Varieties and Agronomy  
     Ato Taye Bizuayehu 
    Inst of Ag Res 
    Areka (Mail only or through 
    IAR headquarters-Addis) 
    Nutrition  
    Dr Zewdie Wolde-Gabriel 
    or Ato Kelbessa 
    Ethiopian Nutrition Inst 
    Addis Ababa Ph 75-15-22 
    Dr. Louis 
    Inter Aid France 
    Bele-Kindo Koshia 
    Ph 15-99-14 
    Culture and Socio-economics 
    Dr Richard Pankhurst 
    and Dessalygn Remato 
    Inst of Ethiopian Studies 
    Addis Ababa Univ Ph 71-29-01 Ext113 
    M/s 
    Awasssa College
     


    Disclaimer

    The designations employed and the presentation of material in this document do not imply the expression of any opinion whatsoever of the UN concerning the legal status of any country, territory, city or area of its authorities, or concerning the delimitation of its frontiers or boundaries.
      


    UN-EUE  Tel.: (251) (1) 51-10-28/29 
    PO Box : 5580  Fax: (251) (1) 51-12-92 
    Addis Ababa, Ethiopia  Email: undp-eue@telecom.net.et