INCIDENCE AND DISTRIBUTION OF DOWNY MILDEW DISEASE ( PERONOSCLEROSPORA SORGHI ) OF SORGHUM IN UGANDA.

Kumi Frank 1,2 , Agbahoungba Symphorien 1 , Badji Arfang 1 , Odong Thomas 1 , Edema Richard 1 , OchwoSsemakula Mildred 1 , Nakubulwa Dianah 1 , Tusiime Geoffrey 1 , Biruma Moses 3 and Rubaihayo Patrick 1 . 1. Department of Agricultural Production, College of Agricultural and Environmental Sciences, Makerere University, P.O. Box 7062, Kampala, Uganda. 2. Department of Crop Science, College of Agricultural and Natural Sciences, University of Cape Coast, P.M.B Cape Coast, Ghana. 3. National Semi-Arid Resources Research Institute (NaSARRI), Serere, Uganda. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History

Sorghum (Sorghum bicolor) ranks as the third most important staple cereal food crop in Uganda after maize and millet. The crop is hampered by several biotic stresses of which Sorghum Downy Mildew disease is one of the most devastating diseases. The disease is caused by a soil-borne fungus Peronosclerospora sorghi but it is also air-borne through conidia. Incidence and severity of downy mildew were conducted between March-June, 2016 and September-December 2016 growing seasons in 13 sorghum-growing districts across ten agroecological zones. Data collected were subjected to analyses of variance, correlation and linear regression. Significant (P < 0.001) differences were recorded for disease incidence and severity across the different agro-ecologies. The mean downy mildew disease incidence varied significantly (P< 0.001) from 49.4 % for Pader to 78.9 % for Namutumba. Disease severity varied significantly (P <0.001) from 2.3 for Pader to 3.5 for Pallisa. Chi-square test for independent association between cropping season and downy mildew disease incidence showed a high significant test statistic values (χ2 = 140.89, d.f = 3, and P < 0.001). A significant (P<0.001) positive correlation was observed between disease incidence and severity (r = 0.498) while a significant negative correlation was observed between disease incidence and rainfall (r = -0.675) due largely to temperature and relative humidity. Multiple linear regression analysis revealed that temperature, relative humidity, rainfall and altitude explained 49 % of the total variation associated to disease incidence. Variations in the incidence and severity of downy mildew disease prompt the need to screen for resistant/tolerant sorghum varieties.
The production of sorghum and its yield are constrained by a myriad of biotic and abiotic factors  with Sorghum Downy Mildew (SDM) (Penorosclerospora sorghi) being one of the most devastating biotic stresses . The disease is characterized by different pathotypes with unique symptoms (Thakur and Pande, 1995). Infected seedlings become stunted and chlorotic with infected young plants likely to die (Jeger, 1998). Older leaves may exhibit alternating parallel stripes of green and yellowish-green to white tissue. Under cool and humid conditions a white downy growth is produced on the lower leaf surface. The chlorotic tissue stripes eventually die and leaves become shredded. Heads produced on these plants may be fully or partially seedless . Globally, SDM affected areas experience significant economic losses as high as 50 % -100 %. These loses occur in tropical lowland, subtropical, mid-altitude, and highland environments when the disease is left uncontrolled (Jeffers et al., 2000). Although empirical information on distribution of SDM in sorghum is limited in Uganda, reports on severity and distribution are available from neighboring countries of Kenya (Ngugi et al., 2002, EPPO, 2014 and Tanzania and Burundi (EPPO, 2014). Limited information and research on the disease coupled with farmers' lack of knowledge in the preventive and control mechanisms of the disease poses a severe threat to the production of the crop. This paper reports the incidence, severity and distribution pattern of SDM disease of sorghum in the different major sorghum growing districts in Uganda.

Methodology:-
Sampling areas and sampling method:-Field surveys were conducted for two seasons (March-June, 2016; major growing season and September-December 2016 minor growing season) for disease incidence and severity of Sorghum Downy Mildew. In all thirteen major sorghum growing districts from ten agro-ecological zones (Wortmann and Eledu, 1999) were surveyed (Table 1). These agro-ecologies are well differentiated by farming systems, soil factors, weather and climatic conditions, altitude and the prevailing vegetation cover. These agro-ecologies and their respective district(s) (Fig 1) were selected purposively (Tongco, 2007) for the economic and livelihood significance derived from sorghum in these areas. Twenty sorghum farms were randomly selected from each district. A total of 10 districts out of the 13 districts were surveyed during the first season (due to drought incidence recorded in Lira, Pader and Serere).

Data collection and analyses:-
A total of four hundred and sixty (460) sorghum farms, were randomly selected and assessed for SDM incidence and severities in the year 2016 for two seasons across ten AEZ in Uganda (Fig 1). Data were collected on the incidence and severity of SDM disease. In addition to the incidence and severity of SDM data, relative humidity, temperature, elevation above sea level and precipitation parameters were taken. Farms were selected at a minimum distance of 500 meters interval apart and the Global Positioning System (GPS) for altitude and co-ordinates were taken at each farm using a GPS receiver. In farmers' fields, SDM incidence was assessed based on percentage as described by Ward et al. (1999). A total of 50 sorghum plants were sampled from each farm and assessed for SDM using key symptoms such as leaf streaks, necrotic leaf, chlorosis, downy appearance, leaf shredding, stunted growth and vein clearing . Plants showing SDM symptoms were counted and expressed as a proportion to the total number of plants assessed. Percentage disease incidence (PDI) was computed using the formula of Ward et al. Field incidence and severity data for SDM from all the surveyed sites were subjected to one-way analysis of variance (ANOVA) (Steel et al., 1997) at a probability level of 5% using Genstat Statistical package version 15. Multiple Linear regression analysis was carried out using Genstat Statistical package. Disease incidence and severity were used as the dependent variables as against temperature, relative humidity, rainfall and altitude which were the independent variables. Arcsine transformation was used to transform SDM incidence scores which were in percentages before carrying out ANOVA. Data for SDM disease incidence and severity means were separated using Fisher's protected Least Significant Difference (LSD) test at P < 0.05 using Genstat statistical package version 15.
Chi-square test for independence or association of incidence and severity data with temperature range was done in Genstat using the maximum likelihood method. Correlation between incidence and severity was done using means. Disease incidence and severity score as well GPS coordinates of the respective sampling locations were used to generate SDM Disease map using Geographic Information Systems (GIS) Arc View 3.2 software.

Results:-
The results for the analysis of associated with downy mildew disease are presented in Table 2. The result showed that, rainfall significantly affected (P < 0.001) the incidence and severity of downy mildew disease. 0.1522 AEZ = Agro-ecological zone; NS = not significant; ***= significant at P < 0.001 and ** = significant at P < 0.01.
The results also showed high significant differences (P < 0.001) of downy mildew disease incidence and severity within Agro-ecological zones and districts. More so, the interaction between District x rainfall for disease incidence was significant (P < 0.001) but the interaction between disease severity for District x rainfall showed no significant difference (P < 0.001).
The results for downy mildew disease incidence and severity for the two growing seasons in 2016 are presented in Fig. 2. The results showed variations in disease incidence and severity across agro-ecological zones. The major season recorded mean downy mildew disease incidence ranged between 75 to 92 % (Fig 2A). Arua district recorded the highest disease incidence (92 %) in the major season, while Kabarole district recorded the lowest disease incidence of 64 % ( Fig. 2A). Kabale, Kumi, Namutumba, Masindi and Pallisa districts all recorded a mean disease incidence greater than 81 % while Nebbi, Hoima and Iganga districts recorded incidence less than 81 % ( Fig. 2A) in the major season.
In the minor season, similar variations for downy mildew disease incidence were recorded. The highest mean disease incidence was 68 % (Fig. 2B), recorded in Namutumba district. Kabarole, Hoima, Lira, Iganga, Pader, and Serere districts all recorded mean disease incidence between 40 and 53 % while Arua, Kabale, Masindi, Kumi, Nebbi and Pallisa districts recorded mean disease incidence between 54 and 67 % in the minor season (Fig. 2B). Similarly, Downy mildew disease severity varied across districts for both seasons (Fig 2). The highest disease severity was recorded in the major season, with a severity score of 4 ( Fig 2C) in Namutumba and Pallisa, while a severity score 4 (severe) was recorded in Namutumba and Pallisa while the remaining eight districts namely; Arua, Hoima, Iganga, Kabale, Kabarole, Kumi, Masindi and Nebbi all recorded a severity score 3 (moderate) in the major season (Fig. 2C). Conversely, the minor season recorded a lower disease severity score (Fig. 2D), with the highest severity score of 3 (moderate) recorded for Arua, Nebbi, Hoima, Masindi, Kabarole, Kabale, Lira, Serere, Kumi, Pallisa and Namutumba while Pader and Iganga districts recorded the lowest severity score in the minor season (Fig.  2D).
Downy mildew disease was found in all the sorghum fields which were surveyed (Table 3). Downy mildew disease was found in all the sorghum fields were the surveyed was carried out (  Mean values with the same alphabet within a column are not significantly different (P < 0.05); CV = coefficient of variation; Lsd = Least significant difference, Std dev = Standard deviation, Temp = Temperature Results of chi-square test for independent association to test the null hypothesis that incidence of downy mildew disease is independent of cropping season are presented in Table 4. The results showed a highly significant test statistic values (χ2 = 225.89, d.f = 3, and P < 0.001). The analysis of variance associated with the multiple linear regression performed between downy mildew disease incidence, weather parameters (relative humidity, rainfall and temperature) and altitude showed a high significant difference (P<0.001) ( Table 5). Similarly, multiple linear regression between downy mildew disease severity, weather parameters (relative humidity, rainfall and temperature) and altitude showed a high significant difference (P<0.001). The associated coefficient of determination (R 2 ) for incidence and severity of downy mildew disease were 0.49 and 0.32 respectively.
Wald test results showed a significant difference (P<0.001) for disease incidence and severity against all the parameters measured (Table 6). The Wald test estimates for disease incidence recorded significant regression coefficient values of -2.26, 0.30, 20.51 and -0.03 for temperature, relative humidity, rainfall and altitude respectively. Severity also recorded a significant regression coefficient values of -0.035, 0.011, -0.446 and -0.001 for temperature, relative humidity, rainfall and altitude respectively.
Correlations were run to compare parameters associated to downy mildew disease and the results showed significant differences (Table 7).  The correlation coefficients showed that downy mildew disease incidence was significantly positively correlated with disease severity (0.498) and also positively correlated to relative humidity (0.445) and altitude (0.058). However, disease incidence was negatively correlated with rainfall (-0.675) and temperature (-0.242). Disease severity was significantly negatively correlated with temperature (-0.109), rainfall (-0.536) and altitude (-0.038) but correlated positively with relative humidity (0.365). Temperature correlated negatively with relative humidity (-0.344) and altitude (-0.635) but correlated positively correlated with rainfall (0.206).

Discussion:-
Downy mildew disease was widely distributed in all the surveyed districts and fields at varying levels of disease incidence and severity and cropping season. In this study, disease incidence and severity recorded in the major season was higher than the incidence and severity recorded in the minor season. This variation could be attributed to the differences in the environmental factors (rainfall, temperature, relative humidity) which prevailed in different seasons as reported by Bigirwa et al., (2000). Bigirwa et al. (1998) and Adipala et al. (1999) reported low prevalence of downy mildew disease with 10% disease incidence (Bigirwa et al. 1998) with isolated distribution pattern of the disease for different cropping seasons. Similarly, low impact of the disease was reported in Nigeria (Ngugi et al., 2002), Zimbabwe, Zambia, Mozambique and Rwanda . Contrary to these earlier reports on the disease distribution and prevalence (EPPO, 2014), this study however showed a significantly high disease incidence and severity pattern with a wider distribution pattern across ten agroecological zones. These variations could be attributed to the different environmental factors in the different agroecological zones (Bigirwa et al. 1998), cropping pattern (Kutama et al., 2010) and growing of susceptible cultivars (Wang et al., 2000).
However, this The results from the study also showed a significant dependent association between disease incidence, severity and cropping season and this explains the variations which were seen recorded for major and minor seasons. This confirms the seasonal variations of the SDM disease as reported by Bock et al., (1998). The study showed significant varying temperatures within the range of 20 -25 °C for the ten different agro-ecological zones, reported 963 Bock et al. (1998) to favour downy mildew conidia production, infection and subsequent epidemic development. High humidity accompanied with with an optimum temperature of 21 -22 °C favours conidia development (Wang et al., 2000). Bock et al. (1998), reported that no conidia sporulation occurred at temperature <13 °C or >30 °C, which therefore explains why the disease was therefore more prevalent in the various growing areas and showed a variation in the disease incidence and severity for the different agro-ecological zones.
Downy mildew disease was found in all the sorghum fields were the surveyed was carried out. Results showed that increased maximum, minimum temperatures and decrease in relative humidity reduced infection of downy mildew disease in Pader, Lira and Nebbi districts. The results further showed that low maximum, minimum temperatures and increased relative humidity favoured infection of downy mildew disease in Hoima, Iganga, Kabale, Kabarole, Kumi, Masindi, Namutumba and Pallisa districts. Infection was also observed to favoured by increased in areas of high precipitation (1011 -1404 mm) and altitude ( >1000 m); this was evident in Kabale, Kabarole, Hoima, Masindi Pallisa and Namutumba districts which recorded both high disease incidence and severity. The study agreed with Kutama et al. (2010), Ngugi et al., (2002) and Bigirwa et al., (1998) who reported high prevalence of SDM in highlands and wetter areas in Nigeria, Kenya and Uganda (Hulluka and Esele, 1992). Previous report by Wang et al., (2000) and Bigirwa et al. (1998) and also showed that SDM is prevalent in highlands associated with cool temperatures with high relative humidity. . Seasonal variations of the disease incidence and severity were also observed in this study and were significantly higher in the major cropping season compared to the minor cropping season due to favourable climatic conditions such as (relative humidity, temperature, rainfall) during the major cropping season as opposed to unfavourable factors (climatic and edaphic) during the minor cropping season (King and Mukuru, 1994). Similar results were reported Bock et al., (1998), that rainfall, suitable night-time temperatures and high relative humidity are required to allow conidia production, infection and subsequent epidemic development the disease. The differences in the disease patterns for agro-ecological zones could also be attributed to other factors which may have contributed to the high disease incidence and severity in the study areas are; mixed cropping which mainly included alternate host plants (maize, sugarcane) and exchange of diseased seeds amongst farmers (Thakur and Mathur, 2002).
There were both positive and negative relationships between weather parameters and altitude on SDM disease incidence and severity. There was a significantly positive disease incidence correlation with disease severity suggests that areas which recorded high disease incidence recorded a similar also in disease severity. The coefficient of determination (R²) explained that 49 % disease incidence and 32 % disease index were associated with weather parameters and altitude. These findings agreed with Barbosa et al. (2006) who reported positive significant relationship between cumulative rainfall, relative humidity, temperature and disease incidence and severity of SDM. More so, significantly negative correlations between maximum temperature and disease incidence and severity suggests high temperature negatively affects the proliferation of the disease infection and epidemic (Sharma et al., 2010) Similar findings were reported by Shawa and Osborne (2011) and Bock et al. (1998) on the response of disease prevalence to favourable climatic conditions. Their finding was that high optimum temperature (20 -23 °C) favours the germination and germ tube growth for disease infection. The trend reflected in this study, were Iganga, Hoima, Kabale, Kabarole, Kumi, Namutumba and Pallisa districts which recorded high relative humidity with a low temperature (<33 °C) and recorded high disease incidence and severity. Wang et al. (2000), reported in Australia and Bock et al. (2002) reported in Zimbabwe that high relative humidity is a predictive determinant factor for downy mildew disease prevalence. This study further found negative significant correlation coefficients between rainfall and incidence and severity of the disease, which suggests that the affect of rainfall had a marked effect in the occurrence and spread of the disease (Thomas, 1992). Kutama et al., (2010) and  also reported that adequate rainfall is needed for conidia production, infection and subsequent epidemic development.
This study showed that SDM was widespread in all the ten agro-ecologies the survey was carried out. Screening for sources of resistance to the disease will be for managing the disease threshold.

Conclusion:-
Downy mildew disease was prevalent and widely distributed in all the surveyed farms at varying incidence and severity levels. Among other factors, weather factors (temperature, relative humidity and rainfall) and altitude were 964 significant in the spread of the disease in sorghum. There is a need for successive assessments of the disease spread in all sorghum growing areas over time and further identify sources of resistance to the disease in Uganda.