+Bioline International Official Site (site up-dated regularly)

Shortage, unavailability and high cost of disease-free seed potato and Ralstonia solanacearum-infested soils have remained key limitations in the management of potato bacterial wilt in Kenya. Attempts to reduce yield losses caused by the disease have been made through on-farm development and promotion of a seed-plot technique which enhances availability of disease-free seed tubers and acts as a window for the introduction of other bacterial wilt management practices to small-holder farmers. Two major potato-producing regions in Kenya, South Kinangop and Meru, which have differences in farmer preferences to potato varieties, disease constraint levels, land use, weather and cropping patterns, were used for evaluation of the seed-plot technique. Certified potato tubers were planted at a spacing of 20 cm by 20 cm in seed plots and at 75 cm by 30 cm in ware plots, the latter simulating farmers' management practices. Data collected included tuber yields and disease incidences. High density planting in seed plots significantly increased land productivity for total tubers per unit area in comparison with low planting density in both experimental regions. In South Kinangop, varieties Tigoni and Roslin Tana produced 2.54 and 2.36 times more tubers per unit area, respectively, in seed plots than in ware plots; in Meru, land productivity was increased 2.08 and 6.75 times for varieties Asante and Kerr's Pink, respectively, planted at seed-plot density. Increases in land productivity were a function of both the planting density and disease pressure. Within five consecutive plantings, bacterial wilt in South Kinangop declined to insignificant levels on farmers' fields. However, seed plots and ware plots alike were infested with the disease in Meru within the first season after planting certified, disease-free seed tubers; in some cases, the disease was introduced by farmers through contaminated feet/shoes as evidenced by occurrence of diseased plants at the seed plot edges.

Key Words: Bacterial wilt control, disease-free tubers, multiplication, seed availability, small-holder farms

La pénerie et l'indisponibilité des semences saines de la pomme de terre et leur coût élevé de même que les sols infestés de Ralstonia solanacearum sont restés des facteurs clés limitant la lutte contre le flétrissement bactérien de la pomme de terre au Kenya. Des tentatives de réduire des pertes de rendement dues à cette maladie ont été faites grâce au développement et la promotion de la technique de production de semences sur des parcelles semencières; laquelle technique augmente la disponibilité de tubercules semenciers sains et sert de créneau à l'introduction d'autres techniques de lutte contre les frétrissements bactériens au niveau des petits paysans. Deux principales régions productrices de la pomme de terre au Kenya, (Kinangop sud et Meru) ont servi pour l'évaluation de la technique de production de semences sur des parcelles semencières. Ces régions englobent des paysans avec des préférences divergentes sur les variétés de pomme de terre et elles ont des niveaux différents de contraintes dues à la maladie de frétrissement, la répartition et utilisation des terres, le climat et le mode de conduite des cultures. Des tubercules certifiés comme sains ont été plantés à 20cm d'interligne et 20cm d'écartement sur des parcelles semencières, et à 75cm x 30cm dans les champs de production, le dernier dispositif simulant les pratiques habituelles des fermiers. Les données collectées incluaient les rendements en tubercules et l'incidence de la maladie. A Kinangop sud, les variétés Tigoni et Roslin Tana ont produit respectivement 2,54 et 2,36 fois plus de tubercules sur les parcelles semencières en comparaison des parcelles en champs de production. A Meru, la production a respectivement augmenté de 2,08 et 6,75 fois pour les variétés Asante et Kerr's Pink, plantées à la densité des parcelles semencières. Dans les deux régions, une plantation à grandes densités sur des parcelles semencières a significativement augmenté la production du nombre total de tubercules par unité de surface, en comparaison des plantations en quantités moins élevées. L'augmentation de la productivité des parcelles semencières était fonction de la densité de plants et de la pression de la maladie. Au Kinangop sud, au cours de cinq plantations consécutives, les frétrissements bactériens ont diminué aux niveaux insignifiants dans les champs de fermiers. A Meru cependant, les semences certifiées saines et plantées en parcelles semencières ont été infestées par la maladie, en première saison; dans certains cas, la maladie étant introduite par les fermiers (pieds/chassures souillés) comme révélé par l'occurrence des plants malades aux bords de parcelles semencières.

Mots Clés: Contôle de frétrissements bactériens, tubercules sains, multiplication, disponibilité de semences, petits fermiers

Shortage or lack of healthy, high-yielding planting materials and the common practice among farmers of planting undersized potato tubers are some of the major limiting factors to production of potatoes (Solanum tuberosum L.) in Kenya (Barton et al., 1997; Kinyae et al., 1994). Consequently, yields have remained persistently below 10t/ha in comparison to over 40t/ha attainable under research station conditions (Lung'aho et al., 1997), despite the area under potato cultivation having expanded almost four-fold from 27,000 ha in 1965 to 96,000 ha in 1995 (Barton et al., 1997). Farmers commonly plant tubers obtained by saving from a previous season's crop or by purchasing from 'neighbours' or market places (Barton et al., 1997). Such tubers have a high potential of harbouring tuber-borne diseases such as bacterial wilt and those caused by viruses. Bacterial wilt (Ralstonia solanacearum) leads to great yield losses (Barton et al., 1997; Michieka, 1993) and reduced store life due to post-harvest rotting (Nyangeri et al., 1984; Eden-Green, 1991; Ajanga, 1993). The disease can only be controlled through integrated management strategies because it defies the use of single strategy approaches when it gets established in the field. However, planting of disease-free seed tubers has been recognised as a very important aspect in boosting potato yields and controlling diseases, bacterial wilt inclusive; other components of integrated management can then be applied more reliably and effectively.

As a way of increasing the availability of satisfactorily healthy seed tubers to resource-poor small-scale farmers, a seed plot technique was developed and introduced in South Kinangop and Meru, both being wilt-infested potato-producing regions of Kenya. These regions experience differences in farmer preferences to potato varieties, disease constraint levels, land use, weather and cropping patterns.

On-farm trials were set up in South Kinangop (Nyandarua District) and Meru (Meru District) regions of Kenya, which had been surveyed earlier and known to have bacterial wilt-infested farms. Six farms were selected in both regions on the basis of presence of bacterial wilt and farmers' willingness to avail land and carry out agronomic practices on the experimental plots. Varieties Tigoni and Roslin Tana were used in South Kinangop while Meru-based experiments utilised varieties Asante and Kerr's Pink, farmer preferences for potato skin colour being the key selection criterion.

Certified seed tubers of each variety were procured from Kenya Agricultural Research Institute's National Potato Research Centre and planted in small plots (seed plots) at a high density (20cm by 20cm spacing) and in ware plots at ware potato (low) density (75cm by 30cm spacing) on each farm. All seed plots measured 2m by 4.5m while ware plots measured 4m by 4.5m in South Kinangop and 6.6m by 7.5m in Meru. Tubers harvested from seed plots were selected and divided into two lots for each variety for subsequent planting: one lot for establishing a new seed plot and the other lot for planting a ware plot in the following season. The experiments were laid out in a randomised complete block design (RCBD) with farms being considered as replicates. Seed plots were ideally established on land presumed to be free from bacterial wilt infestation (land with no history of a solanaceous crop for at least three years) while ware plots were established on regularly cultivated land. The participating farmers were trained on seed plot planting and management as a way of enhancing the separation of seed potato production from ware production; management of ware plots was left to the farmers' discretion. Chemical control of late blight, early blight and arthropod pests was encouraged.

The trials were carried out for five consecutive seasons in South Kinangop while only one season (coinciding with the fifth season in South Kinangop) has been completed in Meru. Data were taken on tuber yields and incidence of bacterial wilt and other diseases. Yields were determined as the number of tubers in five size classes, defined on the basis of tuber diameters, viz. <25 mm, 25-35 mm, 35-45 mm, 45-55 mm and >55 mm; only tubers within the middle three size classes are considered as seed-size tubers for planting. Bacterial wilt was assessed by counting the number of wilted plants in the experimental plots and harvested tubers were further analysed for latent infection by enzyme-linked immuno-sorbent assay on nitrocellulose membranes, NCM-ELISA (Priou et al., 1999).

South Kinangop experiments. Results from the third, fourth and fifth seasons, corresponding to 1998 short rains (September to December), 1999 long rains (April to August) and 1999 short rains (September to December), respectively are presented here. The first two seasons were preliminary trials to evolve working protocols because the seed plot technique had not been tried before.

Land productivity and tuber multiplication. Land productivity for total number of tubers was significantly higher under seed planting density (close spacing) than under ware planting density (wide spacing) for both varieties Tigoni and Roslin Tana over all the three seasons, with variety Tigoni out-yielding Roslin Tana (Table 1). The ratios of tuber yield (tubers/m²) under seed density to that of tuber yield under ware density ranged from 2.31 for var. Tigoni during 1999 long rains to 3.94 for Roslin Tana during 1998 short rains. Consequently, the proportion of land under potato cultivation required for seed production was significantly reduced by planting at seed density, the land requirement being between 2.66 and 3.97 times less than under ware density planting (Table 1). Generally, 15-30% of potato land was needed for seed production under a 75 cm x 30 cm spacing compared to a mere 5-7% under a 20cm x 20cm spacing; the proportions of land required for seed multiplication were lower for var. Tigoni than for var. Roslin Tana (Table 1).

Although the plot size for ware density planting (18m²) was double that of seed density planting (9m²), seed-size tubers obtained from seed plots were far more numerous than those from ware plots over all the three seasons (Table 1). The proportions (%) of seed-size and marketable tubers from both seed and ware plots were fairly similar, with only a few exceptions (Table 1).

Bacterial wilt and other diseases. Only one of the six experimental farm sites experienced bacterial wilt in experimental plots; the disease was recorded in seed plots of var. Tigoni at 2.3% (5 out of 220 plants), 2.3% (5 out of 220 plants) and 1.4% (3 out of 212 plants) during the three consecutive seasons, respectively. Laboratory analysis by NCM-ELISA on tubers sampled from the experimental plots revealed latent infection by Ralstonia solanacearum. Generally, a higher incidence of bacterial wilt was recorded in sampled tubers following NCM-ELISA than by counting wilted plants in the experimental plots (Table 2). Disease incidence was higher during the long rains of 1999 than during the short rains of 1998 and 1999. Bacterial wilt was neither observed in var. Roslin Tana plots nor detected by NCM-ELISA in tubers sampled from the same plots (Table 2).

Routine use of pesticides resulted in adequate control of foliar diseases and arthropod pests in both seed and ware plots. Fusarium dry rot (Fusarium solani) and Black scurf (Rhizoctonia solani) were occasionally observed in some farm sites but at very low incidences.

Meru experiments. Land productivity under the two planting densities in Meru followed a trend similar to that in South Kinangop. The number of tubers produced per unit area of land was significantly higher under seed than under ware planting density for both varieties Asante and Kerr's Pink, the latter variety producing more numerous tubers than the former (Table 3). Increased land productivity under close spacing (seed density planting) resulted in less land requirement for seed production than under wide (ware) spacing (Table 3), as was the case in South Kinangop.

Both seed plots and ware plots suffered bacterial wilt attacks in some farms, although disease-free, certified tubers were planted. However, a higher incidence of bacterial wilt was recorded in ware plots, leading to a wider gap in productivity between seed plots and ware plots. This is reflected clearly in variety Kerr's Pink where a mean bacterial wilt incidence of 9.14% by visual assessment led to a seed density to ware density land requirement ratio of 6.86 (Table 3). With a mean disease incidence of about 1% for variety Asante, the percentage potato land required for seed multiplication was about twice as much under ware density as was under seed density (Table 3). One of the farm sites had a wilt incidence of 95.9% (211 wilted out of 220 plants) in Kerr's Pink ware plot; all the resulting tubers from that plot were also infected by Ralstonia solanacearum. The plot had been established on land previously under potato, although the immediate preceding crop was maize. Other experimental plots on that farm did not experience any bacterial wilt infections; they were established on previously fallow land with no history of potato cultivation.

Some farms, however, did not have bacterial wilt in the experimental plots, despite the occurrence of the disease in other parts of the farms. In cases where the disease occurred in seed plots, symptoms appeared at the periphery or from only one side of the plots.

The proportions (%) of seed-size and marketable tubers were relatively high and similar under both seed and ware planting densities (Table 3).

Land productivity was greatly increased by planting potato seed tubers at a close spacing of 20 cm by 20 cm in seed plots, corroborating earlier findings (Kinyua et al., 1998). This implies that smaller proportions of potato land would be needed for seed multiplication and reflects an increased ease of practising crop rotation in small-scale potato production systems. In return, this flexibility allows a higher possibility of adopting crop rotation as a component of integrated management of bacterial wilt and other soil-borne diseases. Production of seed tubers in seed plots also enhances the separation of on-farm ware from seed production because farmers can pay more attention to the plots, borrowing from their experience of raising vegetable seedlings in nursery beds.

The 5-7% potato land requirement for seed production under the seed plot system reflects a more than 50% land saving in comparison to the 15-30% land requirement under a ware planting system. The land 'saved' should be reserved for the establishment of future seed plots. The land should ideally be maintained as a fallow or planted to crops that are non-hosts to Ralstonia solanacearum so that rotation of seed plots can be practised as one of the measures to reduce soil-borne inoculum, where it exists.

Occurrence of bacterial wilt at the periphery and from one side of seed plots in Meru indicates the possibility of disease spread through infested feet/shoes and farm tools or contaminated runoff water. Therefore, training of farmers on other aspects of integrated disease management, to incorporate field sanitation/hygiene, is an important consideration if the seed plot technique is to realise its potential impact of increasing the availability of disease-free tubers.

Seed multiplication plots should be located on bacterial wilt-free land to avoid possible spread of the disease to non-infested ware production fields. In deed, planting bacterial wilt-free seed, coupled with other management practices, has already been reported to contribute to significant increases in potato yields in many places (Berrios and Rubirigi, 1993; Saumtally, 1993; French, 1994; Shekhawat, 1995; Lemaga et al., 1997). However, the significance of soil-borne inoculum should not be underestimated as displayed by the disastrous effect of the disease on plants raised from disease-free, certified tubers of var. Kerr's Pink on one of the farms. This emphasises the fact that bacterial wilt-free tubers cannot singly solve the disease problem in the absence of other control measures, especially in heavily infested soils and when conditions are ideal for bacterial wilt development.

By employing the NCM-ELISA kit, tuber latent infection levels reflected a higher incidence of bacterial wilt than the incidence recorded by visual assessment; this points to the importance of adopting more sensitive detection methods for R. solanacearum, especially in potato seed production systems. Improved detection of this pathogen would contribute greatly to the quality parameters of seed tubers for bacterial wilt management and, in general, seed potato marketing.

The authors acknowledge the facilitation of research activities by the Kenya Agricultural Research Institute (KARI), CABI Bioscience and International Potato Centre (CIP). Financial support from the Department for International Development (DFID), United Kingdom and PRAPACE Network is also highly appreciated. All the farmers who participated in the on-farm trials are also greatly acknowledged.

TABLE 1. Land productivity and potato tuber multiplication under seed and ware planting densities in South Kinangop^a

Variety	Planting density	1998 short rains						1999 long rains						1999 short rains
Variety	Planting density	Land productivity (tubers/m²)	No. of seed size tubers	% seed size tubers	% potato land for seed production	% marketable tubers	Land requirement ratio*	Land productivity (tubers/m²)	No. of seed size tubers	% seed size tubers	% potato land for seed production	% marketable tubers	Land requirement ratio*	Land productivity (tubers/m²)	No. of seed size tubers	% seed size tubers	% potato land for seed production	% marketable tubers	Land requirement ratio*
Tigoni	Seed	88.8	635.7	80.4	6.4	83.9		103.5	768.8	82.2	5.3	90.6		90.7	644.8	78.4	6.9	91.6
	Ware	29.0	414.0	80.7	24.5	86.3	3.83	44.8	548.5	68.5	15.5	92.6	2.92	35.7	500.5	69.7	19.3	93.0	2.80
Roslin Tana	Seed	93.8	555.5	66.7	7.6	66.8		93.4	657.8	77.7	6.3	79.0		87.6	667.3	84.6b	6.1	86.5
	Ware	23.8	357.5	81.8	30.2	83.0	3.97	30.5	454.2	83.0	18.3	90.1	2.90	37.1	523.5	76.0	16.2	89.9	2.66
LSD0.05		17.4	195.2	13.8	13.6	12.8		13.1	154.4	9.2	3.4	4.2		16.6	163.7	11.6	6.3	5.2
CV(%)		21.8	30.7	14.3	69.1	12.9		14.9	19.6	9.4	23.7	3.7		20.9	22.8	12.3	42.0	4.7

^aAll values are means of six replicates (farms)
^bLand requirement ratio is the ratio of the proportion of land (%) required for seed tuber production under ware planting density to that required under seed planting density

TABLE 2. Incidence of bacterial wilt and latent infection levels under seed and ware planting densities in one farm* in South Kinangop

Variety	Planting density	1998 Short rains		1999 Long rains		1999 Short rains
Variety	Planting density	Field incidence (%)	Latent infection level (%)	Field incidence (%)	Latent infection level (%)	Field incidence (%)	Latent infection level (%)
Tigoni	Seed	2.3	4.2	2.3	15.0	1.4	2.4
	Ware	1.3	1.8	2.6	13.0	1.3	4.5
Roslin Tana	Seed	0.0	0.0	0.0	0.0	0.0	0.0
	Ware	0.0	0.0	0.0	0.0	0.0	0.0

*The farm has a total land of 1 acre and experimental plots were located on low-lying land with a previous history of bacterial wilt on the upper side; an appropriate site selection was difficult

TABLE 3. Land productivity and bacterial wilt incidence under seed and ware planting densities in Meru (1999 short rains)^a

Variety density	Planting	Land productivity						Bacterial wilt incidence
Variety density	Planting	Land productivity (tubers/m²)	No. of seed size tubers	% seed size tubers	% potato land for seed production	% marketable tubers	Land requirement ratio^b	Field incidence (%)	Latent infection level (%)
Asante	Seed	77.2	557.7	79.4	7.6	94.9		1.14 (0-3.18)	2.1
	Ware	37.7	1467.3	78.5	15.8	91.8	2.08	1.38 (0-5.45)	3.7
Kerr’s Pink	Seed	125.0	958.3	85.0	4.2	87.4		1.22 (0-6.82)	4.1
	Ware	46.6	1879.3	81.3	28.8	83.5	6.86	9.14 (0-95.9)	29.6
LSD0.05		21.26	628.9	6.22	4.29	3.58
CV(%)		24.3	42.05	6.24	49.6	3.25