Article body

Introduction

Annual costs for controlling potato (Solanum tuberosum L.) late blight caused by the oomycete Phytophthora infestans (Mont.) de Bary are estimated to be in excess of US$3.2 billion in developing countries (Anonymous 1998). In Kenya, potato production is concentrated in the densely populated highlands (1500-2500 m above sea level), and yield losses attributed to late blight range from 30 to 60% (Lung’aho et al. 2006; Nyankanga et al. 2004; Olanya et al. 2001). Although resistant cultivars can be effective for late blight management, inadequate cultivar deployment and ineffective fungicide application strategies can often result in huge losses and financial catastrophe (Lung’aho et al. 2006).

The management of late blight in Kenya remains a major production challenge because of the continuous potato cultivation that ensures abundant P. infestans inoculum for disease development throughout the year (Nyankanga et al. 2007; Olanya et al. 2001). Foliar blight and tuber blight have frequently been reported to occur and cause considerable losses in the tropical highlands (Nyankanga et al. 2007). Attempts to control late blight are almost entirely done with the use of protectant or systemic fungicides, with little regard to application strategy, in addition to other disease management options (Nyankanga et al. 2004, 2008; Ojiambo et al. 2001). A premix of Ridomil[1] (a.i. metalaxyl and mancozeb) and Dithane M-45 (a.i. mancozeb) has been widely used by potato growers for the control of late blight in the tropical highlands (Nyankanga et al. 2004; Ojiambo et al. 2001).

Published research results in other parts of the world have shown evidence of changes in the population of P. infestans with the emergence of more virulent and metalaxyl-resistant strains (Davidse et al. 1981; Dowley and O’ Sullivan 1981; Fry and Goodwin 1997; Fry et al. 1993). In Kenya, there is no evidence for the presence of new lineages or genotypes of the pathogen, and the P. infestans population consists of US-1 genotype, A1 mating type (Olanya et al. 2001; Vega-Sanchez et al. 2000). However, isolates of P. infestans collected from potato fields in Kenya have shown a high level of metalaxyl insensitivity in areas where growers routinely and excessively use systemic fungicides for blight control in potato and tomato hosts (Hohl 1998; Ojiambo et al. 2001). Metalaxyl insensitivity has also been reported among some isolates in production regions with similar agroecological conditions and production techniques as those used in the Kenyan highlands (Mukalazi et al. 2001).

Even though fungicides are usually more effective when applied before infection occurs, this does not always happen because other factors such as rain, wind, lack of equipment and a precarious financial situation may preclude application or rapid canopy growth. In Kenya, most farmers only apply fungicides after seeing symptoms (Nyankanga et al. 2004). Moreover, poor protectant fungicide applications may not thoroughly cover all foliage in the canopy, resulting in lesions in parts of the canopy. Due to poor fungicide coverage, lesions at various stages of development may exist in a field before fungicides are applied. Compounds with protectant, systemic activity and post-infection efficacy may be very useful in the tropical highlands where there is continuous inoculum and control occurs mostly after infection. However, little documentation is available on the efficacy of fungicide mixtures in this region. Fungicide combinations (systemic + protectant) have been used for late blight management in other parts of the world (Platt 1983, 1985; Samoucha and Gisu 1987). Some of these fungicide mixtures include Tattoo C (31% propamocarb hydrochloride + 31% chlorothalonil), Curzate M8 (8% cymoxanil + 64% mancozeb), Acrobat MZ (9% dimethomorph + 60% mancozeb), Sereno (10 %, fenamidone + 50 % mancozeb) and Noblite (fenamidone + mancozeb) (Inglis et al. 1998; Mayton et al. 2001; Schwinn and Staub 1995). However, the efficacy of these compounds has been shown to be affected by temperature and to vary from region to region (Inglis et al. 1998; Mayton et al. 2001). Moreover, the higher cost of these mixtures and systemic fungicides requires some investigation into economical application rates and their fit into existing spray regimes. In southwestern Uganda, a fungicide application strategy for late blight management based on host resistance and spray scheduling has been investigated and shown to be effective (Kankwatsa et al. 2003).

Because of the occurrence of insensitivity to metalaxyl among isolates of P. infestans and post-infection application of fungicides in the tropical highlands, assessment of the efficacy of other fungicides for the management of late blight in the highland tropics is paramount. Optimization of fungicide application and assessment of fungicide efficacy may improve late blight management, minimize build-up of fungicide resistance and reduce potato production costs. Therefore, the objectives of this study were to (i) evaluate the efficacy of fenamidone + mancozeb and propamocarb HCL + mancozeb mixtures and applications rates on the development and management of late blight in comparison to mancozeb, metalaxyl and an untreated control, and (ii) determine the relationship between fungicide combination and application rates and potato tuber and marketable yield under tropical conditions. Some common anti-oomycete molecules such as cymoxanil, dimethomorph, zoxamide, triphenyltin hydroxide, cyazofamid and strobilurin compounds could not be included in this experiment due to registration requirements.

Materials and methods

Experimental design and layout

The experiments were conducted at four field sites in Kenya: National Potato Centre Tigoni, Mt. Kenya, Kinangop and Molo. The sites were chosen because of their different temperature and precipitation conditions and because differences were observed in late blight pressure among the locations (Ojiambo et al. 2001). The experiments were conducted in 3 m x 4.5 m (width x length) plots consisting of 5 rows planted with 16 seed tubers in a randomized block design. A standard agronomic inter-row spacing of 0.75 m and intra-row spacing of 0.3 m for this region were used for all trials. In all experiments, Kerr’s Pink, a cultivar susceptible to late blight under Kenyan conditions and grown in many regions producing potato for fresh consumption, was used. Sprouted potato seeds were planted on October 20, 1999 and October 19, 2000 during the short rainy season, and on March 15, 2000 during the long rainy season. Diammonium phosphate (DAP) fertilizer was applied at a rate of 500 kg ha-1 and was incorporated into field plots at planting time during all cropping seasons. The plots were hilled three times during crop growth. Limited supplemental sprinkler irrigation was used where necessary and weed management was carried out by conventional tillage between rows.

Measurement of weather variables

Weather data were obtained from a weather station located at the experimental site of Tigoni during both years. Precipitation was recorded with a rain gauge, and temperature and relative humidity were recorded with minimum and maximum thermometers at Tigoni (Table 1). No climatic data was recorded from the Molo, Kinangop and Mt. Kenya locations.

Table 1

Average temperature, relative humidity, precipitation, and number of rainy days recorded at Tigoni, Kenya, during the cropping months in 1999-2000 and 2000-2001

Average temperature, relative humidity, precipitation, and number of rainy days recorded at Tigoni, Kenya, during the cropping months in 1999-2000 and 2000-2001

a Potato plots were planted on October 20, 1999 and October 19, 2000 during the short rainy seasons, and on March 15, 2000 during the long rainy season.

-> See the list of tables

Experimental treatments

Effects of fungicide mixtures and application rates on late blight development at Tigoni

Field trials were conducted at the National Potato Research Center during the 1999-2000 and 2000-2001 cropping seasons. The fungicides used in this experiment were propamocarb HCL (Tattoo, Agr-Evo Products, Inc., USA); mancozeb (Dupont Agricultural Products, Delaware, USA); fenamidone (Noblite, Rhone- Poulenc Products, France), and metalaxyl (Ridomil MZ, CIBA Products Inc., Switzerland). The treatments consisted of propamocarb HCL + mancozeb (at three application rates), fenamidone + mancozeb (at three application rates), mancozeb (at two application rates), metalaxyl and an unprotected control. The experimental plots were arranged in a randomized complete block design with four replications. In all cases, late blight was initiated from natural inoculum during the cropping season. Fungicide applications were initiated at the time of appearance of late blight symptoms on potato plants in experimental plots, similarly to the application strategies used by farmers. Details on application dosage and rates are shown in Table 2. Fungicides were applied at 10-d intervals using a Neptune knapsack sprayer with a hollow cone nozzle. The sprayer was calibrated prior to the first fungicide application so as to deliver a spray volume of 400 L ha-1 at a pressure of 3 bars. Potato plants in the control plots were sprayed with water.

Table 2

Fungicide treatment compound mixtures, rates, dosage and number of applications evaluated for late blight (Phytophthora infestans) control on potato during the 1999-2000 and 2000-2001 cropping seasons at Tigoni, Kenya

Fungicide treatment compound mixtures, rates, dosage and number of applications evaluated for late blight (Phytophthora infestans) control on potato during the 1999-2000 and 2000-2001 cropping seasons at Tigoni, Kenya

a Combinations of fungicide active ingredients used for late blight control.

b Dosage of the active ingredients in the mixtures.

c Number of applications during a cropping season.

d No fungicide applied, plants treated with water.

-> See the list of tables

The experiment was repeated during the 2000-2001 cropping season using the same treatments and number of replications described above. Treatments and plots were randomized between years. The cultivar Kerr’s Pink was used for all experiments.

Effects of fungicide mixtures, application sequences and intervals on late blight development at the Molo, Kinangop and Mt. Kenya locations

Trials using different combinations and sequences of fungicides were conducted at Molo, Kinangop and Mt. Kenya. They were compared with the application strategies used by potato growers obtained from a survey conducted in the three main potato growing regions in the highlands and with the untreated control. The experiment was designed to test the various application strategies with the cultivar Kerr’s Pink at each of the above sites during the 2000-2001 cropping season in relation to farmers’ practices. At all locations, late blight was initiated from natural inoculum during the cropping season. Fungicide applications were initiated at the onset of late blight symptoms. Details on the fungicide compounds, rates, intervals, number and application sequences are summarized in Table 3.

Table 3

Fungicide combinations, application sequences, rates, number of applications and application intervals used in the management of potato late blight (Phytophthora infestans) at the Molo, Kinangop and Mt. Kenya locations during the 2000-2001 cropping season

Fungicide combinations, application sequences, rates, number of applications and application intervals used in the management of potato late blight (Phytophthora infestans) at the Molo, Kinangop and Mt. Kenya locations during the 2000-2001 cropping season

a Different fungicide sequences or treatments used at the three locations.

b Number of applications for each product in a sequence.

c Interval at which fungicides were applied.

d Fungicide application strategy commonly used by potato farmers.

-> See the list of tables

Late blight and tuber yield assessments

At the Tigoni research site, potato late blight severity was assessed by visual rating of foliage in the three inner rows. Visual rating of percent leaf area blighted was performed using a scale of 0 to 100%, where 0% = no disease and 100% = total foliage damaged (James 1971). Disease severity assessment was initiated when the first symptoms were observed in the control plots. Subsequent evaluations of late blight severity were performed at 7-d intervals.

At the highland sites of Molo, Kinangop and Mt. Kenya, field plots were assessed when late blight severity in the control plots exceeded 5% of foliar damage at 7-d intervals.

Tubers from the three inner rows of each experimental plot were harvested. The tubers were washed and their weight was determined (Salter suspended weighing machine, Model 235). Tubers in all plots were also separated into marketable (tuber diam > 6.4 cm) and unmarketable (tuber diam < 6.4 cm). Tuber weight was converted to metric tons per hectare for subsequent analysis.

Statistical analysis

Foliar blight was analyzed by using disease severity assessment to compute the area under the disease progress curve (AUDPC) as described by Shaner and Finney (1977).

The effects of fungicide mixtures and application rates on foliar blight severity, AUDPC, and total and marketable yield were computed by ANOVA using SAS Proc GLM analysis (SAS /STATS Version 9, Cary, NC). For the Tigoni experiments, the effects of fungicide mixtures and application rates were analyzed as factorial combinations. When ANOVA indicated significant differences among treatments, mean late blight severity and total and marketable tuber yield were separated by Fisher’s LSD statistics. The progress of late blight severity among treatments and across assessment periods was plotted and examined graphically.

Results

Weather conditions

Weather conditions during the 1999-2000 cropping season were favourable to late blight development. Total precipitation in November and December was fairly high and relative humidity was also sufficient for late blight infection (Table 1). In the second rainy season (mid-March to July in 2000), precipitation was unreliable. The crop was brought up under sprinkler irrigation, but no late blight symptoms were observed and no data were obtained. During the 2000-2001 cropping season, temperature and precipitation were sufficiently conducive to pathogen infection, though average relative humidity was somewhat lower (Table 1). No climatic data were recorded at the Molo, Kinangop and Mt. Kenya locations; however, those regions are wet and cool due to the high altitudes.

Effects of fungicide mixtures and application rates on late blight development at Tigoni

During the 1999-2000 season, the first late blight symptoms were observed 41 d after planting. Disease increased rapidly in unprotected plots with 98.5% late blight severity 35 d after initial detection of the disease. All fungicides, regardless of the application rate, had significantly lower AUDPC and final disease severity scores than the untreated control (Table 4). The propamocarb HCL + mancozeb application rate of 4 L ha-1 resulted in significantly lower late blight severity than application rates of 2 L ha-1 and 3 L ha-1 (Table 4). Late blight severity was significantly lower (P < 0.05) in propamocarb HCL + mancozeb treated plots compared with the ones treated with mancozeb (Tables 4 and 5). Variations in disease development and severity were observed among treatments at various assessment dates. Potato plants treated with the propamocarb HCL + mancozeb mixture showed moderate disease development, while mancozeb-treated plants exhibited higher late blight severity and metalaxyl-treated plants had the lowest level of disease severity at each assessment (Fig. 1). No significant differences in final late blight severity among potato plants treated with fenamidone + mancozeb were observed for the various application rates (Table 4). The application of metalaxyl resulted in significantly (P < 0.05) lower late blight severity compared with all other treatments (Table 4 and Fig. 1). During the 1999-2000 growing season, the fungicide mixtures had variable suppressive effects on disease development. Orthogonal contrasts showed that fenamidone + mancozeb versus the mancozeb treatment was not significant, while propamocarb HCL + mancozeb versus the mancozeb treatment was significant (Table 5).

Table 4

Effect of fungicide mixtures on late blight (Phytophthora infestans) severity, AUDPC, and mean total and marketable tuber yield during the 1999-2000 and 2000-2001 cropping seasons at Tigoni, Kenya

Effect of fungicide mixtures on late blight (Phytophthora infestans) severity, AUDPC, and mean total and marketable tuber yield during the 1999-2000 and 2000-2001 cropping seasons at Tigoni, Kenya

a Fungicides and fungicide mixtures used in efficacy studies.

b AUDPC calculation based on visual disease assessments (%) over time.

c LSD values refer to a comparison of means for the different fungicides within one cropping season.

d LSD values in parentheses refer to a comparison of means for the different rates of application of one product or mixture.

-> See the list of tables

Table 5

Orthogonal contrasts of mean final disease severity, AUDPC, and total yield of potato in field experiment conducted with fungicide mixtures during the 1999-2000 and 2000-2001 cropping seasons at Tigoni, Kenya

Orthogonal contrasts of mean final disease severity, AUDPC, and total yield of potato in field experiment conducted with fungicide mixtures during the 1999-2000 and 2000-2001 cropping seasons at Tigoni, Kenya

*,** mean significantly different at 5% and 1%, respectively.

-> See the list of tables

Figure 1

Progress of potato late blight (Phytophthora infestans) on cultivar Kerr's Pink treated with different fungicides at the Tigoni Research Station during the 1999-2000 cropping season.

Progress of potato late blight (Phytophthora infestans) on cultivar Kerr's Pink treated with different fungicides at the Tigoni Research Station during the 1999-2000 cropping season.

On each day of assessment, values = mean disease severity. The vertical bars represent standard errors of the mean.

-> See the list of figures

At Tigoni, disease severity was considerably higher during the 2000-2001 cropping season compared with the 1999-2000 season (Table 4 and Fig. 2). During the final assessment, the control and propamocarb HCL + mancozeb (4 L ha-1) treated potato plants had the highest and lowest late blight levels, respectively (Table 4). Late blight severity in plants treated with propamocarb HCL + mancozeb applied at rates of 2 L ha-1 (10.6%) and 4 L ha-1 (4.5%) differed significantly (P < 0.05). Plants treated with mancozeb (1.5 kg ha-1 and 2.5 kg ha-1) and metalaxyl (2.5 kg ha-1) had a final late blight severity of 63.5, 51.4 and 8%, respectively. Development of late blight was greater in the control plants compared with plants treated with mancozeb, fenamidone + mancozeb and metalaxyl (Fig. 2). Development of late blight was also greater in the control plots compared with the ones treated with propamocarb HCL + mancozeb. Orthogonal contrasts of fungicide treatments versus unprotected control, and of mixture treatment comparisons, revealed significant differences in late blight severity, AUDPC and total yield between many treatments (Table 5).

Figure 2

Progress of potato late blight (Phytophthora infestans) on cultivar Kerr's Pink treated with different fungicides at the Tigoni Research Station during the 2000-2001 cropping season.

Progress of potato late blight (Phytophthora infestans) on cultivar Kerr's Pink treated with different fungicides at the Tigoni Research Station during the 2000-2001 cropping season.

On each day of assessment, values = mean disease severity. The vertical bars represent standard errors of the mean.

-> See the list of figures

Effect of fungicide mixtures, application sequences and intervals on late blight development at the Molo, Kinangop and Mt. Kenya locations

All fungicide treatments, regardless of the fungicide mixture, sequence and application rate, had significantly lower disease severity and AUDPC values than the untreated control (Table 6).

Table 6

Effect of different fungicide mixture treatments and application sequences on potato late blight (Phytophthora infestans) severity and tuber yield at three locations in the Kenya highlands during the 2000-2001 cropping season

Effect of different fungicide mixture treatments and application sequences on potato late blight (Phytophthora infestans) severity and tuber yield at three locations in the Kenya highlands during the 2000-2001 cropping season

a Molo: sequences 1 and 2 = mancozeb (applied twice) and fenamidone + mancozeb and propamocarb HCL + mancozeb (applied once); sequence 3 = two applications of mancozeb and two applications of metalaxyl. See Table 3 for application rates.

b Kinangop: sequence 1 = mancozeb, propamocarb HCl + mancozeb, fenamidone + mancozeb, propamocarb HCL + mancozeb, applied in that order; sequence 2 = propamocarb HCL + mancozeb, mancozeb, propamocarb HCL + mancozeb (twice), applied in that order; sequence 3 = metalaxyl applied four times during a season. See Table 3 for application rates.

c Mt. Kenya: sequence 1 = mancozeb, fenamidone + mancozeb, propamocarb HCL + mancozeb, each applied twice in that order; sequence 2 = mancozeb (twice) and propamocarb HCL + mancozeb, fenamidone + mancozeb, propamocarb HCL + mancozeb, fenamidone + mancozeb, in that order; sequence 3 = six and three applications of mancozeb and metalaxyl, respectively. See Table 3 for application rates.

d Fungicide application strategy commonly used by potato farmers.

-> See the list of tables

In Molo, fungicide mixtures and application sequences significantly reduced late blight severity compared with the untreated control (Table 6). However, no significant differences were observed in late blight severity among treatments (sequences) 1, 2 and 3 at this location. In this region, two applications of mancozeb (2.5 kg ha-1) followed by one application of fenamidone + mancozeb (1 kg ha-1) and a single application of propamocarb HCL + mancozeb (4 L ha-1) at 7-d intervals (sequence 1) did not significantly improve late blight control compared with sequences 2 and 3.

At the Kinangop and Mt. Kenya locations, late blight levels were also significantly higher in untreated plots compared with treated ones (Table 6). In these two regions, treatments (application sequences and combinations of spray intervals) were not significantly different, except when compared with the untreated plants (Table 6).

Effect of fungicide mixtures and application rates on tuber yield at Tigoni

During the 1999-2000 season, propamocarb HCL + mancozeb and mancozeb alone, both applied at different rates, resulted in significantly higher yield (P < 0.05) than that obtained with the untreated plants (Table 4). Application of propamocarb HCL + mancozeb at a rate of 4 L ha-1 resulted in significantly higher yields compared with application rates of 2 and 3 L ha-1. Plants treated with metalaxyl had the highest total yield (36 t ha-1) during the 1999-2000 season. Mancozeb treatments at 1.5 and 2.5 kg ha-1 resulted in total tuber yields of 14.1 and 16.7 t ha-1, respectively. Marketable tuber yield also differed significantly between all fungicide treatments and the control (Table 4). Marketable yields ranged from 2 t ha-1 (untreated) to 26 t ha-1 (metalaxyl-treated plants) in 1999-2000. No significant differences in yield were observed between fenamidone + mancozeb and mancozeb alone. The metalaxyl-treated plants showed significantly higher marketable tuber yield than all other fungicide-treated plants in 1999-2000.

In 2000-2001, total yield ranged from 12 to 21.9 t ha-1 in the untreated control and metalaxyl-treated plants, respectively (Table 4). No difference in total yield was recorded among treatments in which three rates of propamocarb HCL + mancozeb were applied. Plants treated with metalaxyl had significantly higher total yield compared with those treated with mancozeb (Tables 4 and 5). Marketable tuber yield ranged from 9.3 t ha-1 in untreated control plots to 17.3 t ha-1 in metalaxyl-treated plots (Table 4). No significant differences in yield were observed between fenamidone + mancozeb and mancozeb alone (Table 5).

Effect of fungicide mixtures, application sequences and intervals on yield at the Molo, Kinangop and Mt. Kenya locations

The three regions showed differences in efficacy of spray programs containing propamocarb HCL + mancozeb and fenamidone + mancozeb (Table 6). In Molo, the sequence involving two applications of mancozeb (2.5 kg ha-1) followed by one application of fenamidone + mancozeb (1 kg ha-1) and one application of propamocarb HCL + mancozeb (4 L ha-1) at 7-d intervals gave numerically better total and marketable yields, even though the results were not statistically significant. In Kinangop, there were no significant differences among the sequences while for the Mt. Kenya region, the best yields were achieved using the third sequence, i.e. mancozeb followed by metalaxyl (Table 6).

Discussion

The mixtures of propamocarb HCL + mancozeb and fenamidone + mancozeb were effective in controlling late blight in established epidemics compared with the application of mancozeb alone or the untreated control. However, none of the fungicide mixtures was more effective than metalaxyl. The relatively enhanced disease control obtained with fungicide mixtures under moderate to severe late blight conditions suggests that fungicide mixtures can provide effective late blight management in established epidemics compared with the use of a protectant fungicide. Fungicide mixtures have been shown to provide effective late blight control incited by the US-1 genotype of P. infestans (Mayton et al. 2001; Platt 1985). Mixtures of systemic and protectant fungicides were equally effective for the control of late blight and downy mildew in various experiments (Samoucha and Gisu 1987). Mixtures have also been shown to be very effective when applied prior to infection (Cianchetti et al. 2000). The propamocarb HCL + mancozeb mixture has been shown to suppress epidemics through the suppression of disease progress and lesion expansion (Mayton et al. 2001). Fenamidone, from the imidazolinone family, possesses preventive and anti-sporulant properties that have an impact on direct germination of sporangia (Bardsley et al. 2002; Mercer and Latorse 2003). In our study, the late blight control exhibited by the mixtures could be attributed to the active ingredients in the mixture which have different chemistries, and their efficacy may be attributable to their different modes of action. Perhaps this may also be due to the fact that more active ingredients were applied in the mixtures, or it may be a result of both the fungicide chemistries and quantity factors.

Metalaxyl-treated plants sustained less late blight damage compared with the plants exposed to fungicide mixtures. In general, metalaxyl is effective for controlling diseases incited by oomycetes because it is absorbed by the leaves and roots under various environmental conditions (Easton and Nagle 1985). This may be the case in our tropical experimental setting where variable climatic conditions exist, but metalaxyl is still effective on late blight incited by the US-1 genotype. The relatively higher late blight disease severity recorded in mancozeb-treated plants at 1.5 and 2.5 kg ha-1 suggests that protectant fungicide treatments initiated after the appearance of late blight symptoms or after disease establishment may not be effective for disease control. This is in contrast to other late blight research reports, which showed that mancozeb applied as a protectant can be effective in reducing the impact of late blight under tropical conditions (Namanda et al. 2004). It is possible that application timing and intervals between applications for mancozeb may not have been sufficient to optimize its effectiveness in this experiment. Therefore, contact or protectant fungicides need to be used preventatively instead of post-infection for effective control of late blight (Johnson et al. 2000; Mayton et al. 2001).

Differences in late blight severity were observed between the 1999-2000 and 2000-2001 cropping seasons; however, the use of mixtures was effective for disease control in both seasons. The differences in disease levels may be attributed to the dry weather conditions that were experienced immediately following the appearance of late blight symptoms (41 d after planting) during the 2000-2001 season or to fungicide efficacy. Variation in late blight incidence and severity has been reported previously for various locations in Kenya (Nyankanga et al. 2004; Olanya et al. 2006). The variation in disease development among the different treatments may have been due to the different chemistries of the fungicides used. Fenamidone possesses the special property of “apical diffusion”, which protects young developing organs from a treated bud (Bardsley et al. 2002; Gisu 2002; Mercer and Latorse 2003; Rhone 2003). As the newly formed plant tissues grow, the product is distributed and protects the young shoots. This property of fenamidone helps to protect the upper part of the potato plant.

Different regions showed differences in effectiveness of spray programs containing propamocarb HCL + mancozeb and fenamidone + mancozeb. However, alternation of fungicide mixtures with either metalaxyl or mancozeb did not improve the efficacy of either foliar blight control or yield. This is contrary to what we expected, especially the alternation of fenamidone + mancozeb and propamocarb HCL + mancozeb, which we anticipated to have better control. In other late blight studies, fungicide application schemes that included cymoxanil, dimethomorph, or propamocarb hydrochloride provided slightly better control of late blight in a North American fungicide trial than schemes that did not include those fungicides (Inglis et al. 1998).

High tuber yield was obtained with plants treated with fungicide mixtures and metalaxyl during both years. This suggests that compound mixtures are effective for blight control and can result in improved tuber yield under humid tropical conditions. This also implies that mixtures of fenamidone + mancozeb and propamocarb HCL + mancozeb can be used as substitutes for metalaxyl in late blight control incited by the US-1 genotype under wet and humid tropical conditions, or when foliar blight is moderate. Higher tuber yields resulting from the use of different fungicide compounds, application rates and strategies have previously been reported in the highland tropics (Kankwatsa et al. 2003; Ojiambo et al. 2001). It should be noted that under our experimental setting, it was difficult to determine whether the effect of mixtures was synergistic or additive due to the fact that propamocarb HCL and fenamidone were not applied alone as treatments. At low to moderate disease levels, there were no differences between the rates; the lower rate could therefore be used. However, rate effects resulted in significant differences in disease levels. We were unable to analyze the costs and benefits associated with fungicide mixtures in this study. This is a drawback because costs related to fungicides, application rates as well as other variables, which would enable us to compute yield revenues, were not compiled.

We conclude that fungicide mixtures can be beneficial to late blight management in established epidemics. Apart from their effective disease control, half-rates of fungicides are applied in the mixtures, which reduces the actual amount of recommended fungicide rates, the costs as well as environmental degradation. Another significant attribute is the possible indirect effect of the mixtures on minimizing selection pressure in pathogens as a result of the multi-site action of mixture compounds. Even though the economic benefit of mixtures was not precisely quantified in this experiment, it is presumed that the use of mixtures can result in better late blight control and acceptable tuber yield, as demonstrated by our results.