This study on the epidemiology of banana bunchy top disease (BBTD) was carried out in the context of small-scale farms in Burundi for an integrated management approach. Banana trials were established in farmers’ fields comparing different plot locations, while spatial and seasonal occurrence of aphid vectors was evaluated at three different altitudes. In addition, serological tests were performed on banana leaf samples to confirm the presence and titre of the virus. The results showed that BBTD incidence varied among banana cultivars and locations. Nine months after plot establishment, BBTD incidence ranged from 21·8% to 56·4% in plots within affected fields, while a range of 0–12·3% was reported in plots located between 5 and 30 m away from affected banana fields. Aphid numbers were highest in the dry season. These aphids were able to acquire and transmit the virus irrespective of altitude. A mean incubation period of 21 and 84 days was observed at low (780 m a.s.l.) and high (2090 m a.s.l.) altitude, respectively. Thus, a holistic approach, taking into account banana cultivar, plot location, disease-free planting material and regular field sanitation, should be promoted for long-term BBTD management.
Banana bunchy top disease (BBTD) is one of the most economically important diseases in many banana-producing areas of Africa, Asia and the South Pacific (Furuya et al., 2005; Hooks et al., 2009). Between 1913 and 1920, the banana-growing industry in Australia was almost completely destroyed by the disease (Magee, 1927; Hooks et al., 2009). In the 1990s, the first severe outbreak of BBTD in Africa was estimated to have reduced banana production in the Nkhatabay and Nkhotakota districts of Malawi from 3500 ha to about 800 ha (Soko et al., 2009; Kumar et al., 2011). Although accurate estimates of yield losses are lacking for the Great Lakes countries of Africa, about 90% yield loss was reported in severely BBTD-infected plants of susceptible cultivars such as Poyo and AAA-Cavendish in a screening trial conducted in the Rusizi valley, in Burundi (Niyongere et al., 2011a). Because of the high destructive potential of the disease, Banana bunchy top virus (BBTV) was listed as one of the world’s 100 worst invasive species and the International Plant Protection Convention included it as a pathogen to be subjected to rigorous quarantine measures (IPPC, 2010; Kumar et al., 2011).
BBTD is dispersed over long distances through the exchange of infected suckers and/or through non-virus indexed tissue culture (TC)-derived plantlets (Thomas & Caruana, 2000). Through these propagules, the virus is introduced into new areas where the banana aphid Pentalonia nigronervosa (Hemiptera, Aphididae) occurs (Almeida et al., 2009). This aphid, which is the known vector of BBTV, transmits the virus while feeding on banana plants.
The aphid was initially reported to have occurred on member species of the Zingiberaceae and Araceae families, but morphological and morphometric studies confirmed that these aphids were in fact a separate species (P. caladii; Bhadra & Agarwala, 2010; Foottit et al., 2010). Pentalonia nigronervosa was, therefore, confirmed to have high host specificity to Musa spp., and has been found in almost all banana-growing countries (Kumar et al., 2011). The life span of an aphid ranges from 19 to 26 days and during this period it will produce up to 20 offspring in optimal conditions of 24–28°C (Völkl et al., 1990; Yasmin et al., 1999). The aphid acquires the virus after at least 4 h of feeding on an infected plant and retains BBTV throughout its adult life but does not transmit it to its progeny (Nelson, 2004). The winged aphids, which often develop after 7–10 generations of wingless individuals, are probably responsible for the spread of the virus (Nelson, 2004; Young & Wright, 2005). The winged aphids can transmit the virus to a healthy banana plant by feeding on it for as little as 15 min to about 2 h (Dale, 1987; Hu et al., 1996). Ambient temperatures above 14°C enable the aphids to fly (Jones et al., 2010). Aphids are most frequently observed near the base of banana plants, in between leaf sheaths, and at the base of the youngest unfurled leaf (Robson et al., 2006). In addition, Young & Wright (2005) reported a spatial edge effect, with larger aphid colonies observed at the edge of plantations. Yellow traps filled with soapy water are generally used to collect and monitor winged aphids (McCartney & Fitt, 1985).
In contrast to other banana diseases, such as xanthomonas wilt (Karamura & Tinzaara, 2009), BBTV cannot be transmitted mechanically through garden tool use (Wardlaw, 1961; Kumar et al., 2011). The incubation period of the disease (i.e. up to the appearance of dark-green leaf streaks on the leaf lamina) varies from 19 days in the summer (25–33°C) to 125 days in the winter (June to August, −2 to 11·5°C) in Australia (Magee, 1927; Allen, 1987). An incubation period of 25–85 days (23–29°C) has also been reported in Hawaii (Hooks et al., 2008). Previous research showed that this incubation period, after screenhouse inoculation, is positively correlated with the age of the host plant and negatively correlated with the number of virulent aphids feeding on the plant (Robson et al., 2006; Hooks et al., 2008). The incubation period is also influenced by banana cultivar, rate of leaf emergence and genetic differences between BBTV strains (Wu & Su, 1989; Robson et al., 2006; Hooks et al., 2008). Madden et al. (2006) defined an epidemic as ‘a change in disease intensity in a host population over time and space’. In addition, McCartney & Fitt (1985) suggested that disease management needs an understanding of host plant (i.e. cultivar, type of planting material, growth), epidemic environment (i.e. temperature, rainfall, wind) and pathogen availability and dispersibility.
The exponential increase of BBTD was reported by Smith et al. (1998) and P. nigronervosa’s high transmission efficiency suggests that disease incidence can increase rapidly in the absence of aphid population control (Hu et al., 1996; Robson et al., 2006). The recommended strategies for controlling BBTD in different regions of the world include identifying and destroying virus-infected plants as early as possible, replanting with virus-free TC plants in isolated fields and controlling aphid-vector populations associated with banana plants year-round irrespective of the presence of the disease (Young & Wright, 2005; Robson et al., 2006; Hooks et al., 2009). BBTD was reported to be more prominent at elevations below 1300 m a.s.l. in eastern Democratic Republic of the Congo (DR Congo) (Walangululu et al., 2010). Although most BBTV-infected mats can be found at the lowest elevations in the Rusizi valley (encompassing parts of Burundi, Rwanda and eastern DR Congo), some infected banana mats were observed up to 1700 m a.s.l. in the hills surrounding the valley (Niyongere et al., 2011b). However, information is lacking on disease epidemiology at lower and higher elevations in central Africa. In-depth knowledge of BBTD epidemiology is required for developing an effective management strategy in the small-scale farming system context of the Great Lakes region of Africa. The purpose of the current study was to understand BBTD epidemiology across different agroecological zones and altitudes in Burundi. On-farm and greenhouse trials were established to assess: (i) BBTD epidemiology at different altitudes; (ii) aphids’ ability to acquire and transmit BBTV at low and high altitudes; and (iii) optimal plot location, taking into account distance from a diseased field and wind direction for establishing new plantations with minimum risk of early reinfection.
Materials and methods
Assessing influence of altitude on BBTD spread at Benga and Kagazi
The possible effect of altitude on BBTD spread was assessed at two BBTD-affected locations in Burundi. The first site was at Benga in Isale commune, Bujumbura Rural western province, located in a hilly region at 1268 m a.s.l., while the second site was at Kagazi, Cibitoke commune, northwestern province, which is a flat valley region 893 m a.s.l. Banana cv. Yangambi Km5 (Musa AAA genome, beer type) is mainly intercropped with beans and maize in Kagazi, while more dense plantations of East African highland banana (AAA-EA, beer type) and Gros Michel (AAA, dessert), in association with fruit trees, dominate in Benga. Host-farm/er selection criteria included: location in an affected area, easy access to selected plots and willingness to collaborate in the trials. Four plots were established at each location and monitored over a 24-month period (December 2007–December 2009).
Each plot had three banana cultivars, namely the cooking cultivar Nyambururu (AAA-EA), and the dessert cultivars Kamaramasenge (AAB) and Grande Naine (AAA). Three tissue-cultured (TC) plantlets and three suckers per cultivar were planted in each plot. Hence, a total of 72 TC plantlets and 72 suckers were established across both sites. These planting materials were obtained from the Institut de Recherche Agronomique et Zootechnique (IRAZ) nursery (where TC plantlets were hardened) and from the IRAZ banana field collection (source of suckers). The IRAZ Mashitsi research station in Gitega, central province is known to be located in a BBTV-free zone. Inter- and intra-row plant spacing was 3 and 2 m, respectively. Trial plots were established at Benga and at Kagazi in sites which were approximately 5–10 m from an existing field containing BBTV-infected plants. Weeding was regularly carried out using a hand hoe, while no organic or inorganic fertilizers were applied to the plants during the trial. Disease incidence was monitored at weekly intervals and monthly average incidence levels were calculated per location.
Preliminary results showed that the distance between existing infected plants and newly established plants influences reinfection rates in newly established plots. Therefore, further studies were undertaken to determine the influence of the distance between a newly established plot and an existing affected banana field on disease spread.
Influence of plot location on disease incidence and aphid population
BBTD spread and aphid populations were assessed for 9 months between April and December 2009 in two rural villages (890–895 m a.s.l.) located along the third and fourth roads of Rugombo village, Cibitoke commune, northwestern province, Burundi. Six treatments were established according to the distance and relative location (south or north, predominantly corresponding to upwind or downwind, respectively) from existing BBTD-affected banana plantations. The treatments comprised plots located: (i) inside an affected banana plantation where infected mats had not been uprooted; (ii) inside an affected banana plantation where infected mats had been uprooted and all plant debris was removed from the field immediately after uprooting; (iii) approximately 5 m outside and north of an affected banana plantation; (iv) 30 m outside and north of an affected banana plantation; (v) approximately 5 m outside and south of an affected banana plantation; and (vi) 30 m outside and south from an affected banana field. The first two treatments are the most common farm practices in the region.
Two plots were established at each of the six locations, and there were thus a total of 12 plots in the trial. In each plot, banana cultivars Americani (AAA, Cavendish), Gisandugu (ABB), Kamaramasenge (AAB) and Yangambi Km5 (AAA) were planted. These cultivars were in the form of TC plantlets sourced from the IRAZ nursery and corresponding suckers were obtained from the banana field collection of IRAZ. In addition, suckers of local cultivar Yangambi Km5 (AAA), sourced from BBTD-free fields at Rugombo, Cibitoke northwestern province, were confirmed to be BBTV-free after enzyme-linked immunosorbent assay (ELISA) testing and were used as a local control for the disease incidence. Each plot measured 15 × 12 m and inter- and intra-row plant spacing was 3 and 2 m, respectively, while planting holes measured 40 × 40 × 40 cm. Three 3-month-old TC plantlets and three suckers of each test cultivar, as well as six suckers of the local control Yangambi Km5, were planted per plot. Thirty-six TC plantlets and 36 suckers per test cultivar plus 72 suckers of the control Yangambi Km5 were planted, making a total of 144 TC plantlets and 216 suckers. The fields were maintained according to farmers’ practices (i.e. no organic or inorganic fertilizer was used; a hand hoe was used for weeding).
BBTD spread was assessed weekly, based on disease symptoms, for a period of 9 months. In addition, aphid colonies were scored on all plants bimonthly throughout the trial. BBTV presence was confirmed using the commercial AGDIA kit for triple antibody sandwich (TAS)-ELISA in plants with symptoms from the six plot locations; a yellow coloration was observed for positive samples. The virus concentration of each sample was determined at 405 nm using a spectrophotometer; an optical density (OD) of 0·1 was considered the limit between positive and negative samples.
Seasonal and spatial distribution of banana aphids
Winged aphid populations were monitored using yellow water traps in order to determine their distribution in three different banana-growing agroecological zones in Burundi. The possible seasonal effect on aphid counts over a 1-year period (January 2009–December 2009) was analysed using available data on temperature and rainfall. Two yellow water traps were positioned within banana plots at each of the three sites, namely: (i) Kagazi, Cibitoke northwestern province (890 m a.s.l.); (ii) Benga, Bujumbura Rural western province (1268 m a.s.l.); and (iii) the Mashitsi-IRAZ regional banana field collection, Gitega central province (1645 m a.s.l.).
The yellow water traps were also placed in a banana plantation at the Mparambo (898 m a.s.l.) ISABU agricultural research station, Rusizi valley, Cibitoke province. Two traps were positioned inside the banana field, while another two traps were positioned southwards at 30 and 50 m away from the banana field to assess the spatial distribution of aphids in and around the banana field. The aphids were collected during two consecutive years (September 2007–September 2009).
At all sites, the trapped winged aphids were collected from traps three times per week on Mondays, Wednesdays and Fridays. These aphids were then preserved in bottles containing 70% ethanol pending formal identification and enumeration using a dissecting microscope at the Entomology laboratory of ISABU, Bujumbura, Burundi.
Assessing the effect of altitude on BBTV transmission rates
The BBTV transmission rates were assessed at two different altitudes in Burundi. The first location was the ISABU headquarters in Bujumbura (780 m a.s.l., with average minimum and maximum temperatures of 17 and 32°C, respectively), while the second location was at the ISABU research station at Gisozi (2090 m a.s.l., with average minimum and maximum temperatures of 6 and 25°C, respectively).
Banana aphids were collected in farmers’ fields at Rugombo, Rusizi valley, Cibitoke, northwestern province (890 m a.s.l., average temperature 25°C), which is highly affected by bunchy top, and at the IRAZ banana field collection at Mashitsi (1645 m a.s.l., average temperature 19°C) in Gitega, central province, located in a BBTD-free zone.
Hardened 3-month-old TC plantlets of cv. Intobe (AAA-EA) were obtained from the IRAZ nursery. Six virus transmission trials were conducted. The first two direct transmissions were carried out: (i) using aphids from IRAZ-Gitega on TC plantlets at Bujumbura to determine whether the Gitega aphids carried the virus; and (ii) using aphids from Cibitoke on TC plantlets at Bujumbura to determine whether the Cibitoke aphids carried the virus. For the indirect virus transmission, aphids were fed in both Bujumbura and Gisozi for 24 h on BBTV-infected plants prior to their introduction to the test plantlets. This experiment was conducted in order to assess whether aphids are able to acquire and transmit the virus at different altitudes. These indirect transmissions were carried out: (iii) using aphids from Cibitoke on TC plantlets at Bujumbura; (iv) using aphids from Cibitoke on TC plantlets at Gisozi; (v) using aphids from IRAZ-Gitega on TC plantlets at Bujumbura; and (vi) using aphids from IRAZ-Gitega on TC plantlets at Gisozi.
After an inoculation access period of 48 h, all aphids used in the virus transmission trials were killed using a systemic insecticide (dimethoate) in order to avoid aphid spread.
Six Intobe TC plantlets were used in each of the trials, giving a total of 24 TC plantlets in four trials carried out at Bujumbura, while 12 TC plantlets were used in two trials conducted at Gisozi. At each location, six control plants (no inoculation) were added. The transmission trials entailed the placing of 20 aphids in between the two youngest leaf petioles of each test TC plantlet, making a total of 720 aphids.
Visible BBTD symptoms and leaf emergence rates were recorded at weekly intervals over a 4-month period (April–July 2009) in order to determine the incubation period at different altitudes. In addition, 36 banana leaf samples were collected on BBTV-infected (i.e. all six leaf samples for each of the five treatments making a total of 30 samples) and control (i.e. six leaf samples; one leaf sample per treatment) TC plantlets for BBTV detection using TAS-ELISA.
Temperature and rainfall data for the period January 2007–December 2009 were obtained from the meteorological stations of the Institut Géographique du Burundi (IGEBU) located at Bujumbura airport, Gitega airport and the Mparambo ISABU research station which are representative for the study sites located in Bujumbura Rural, Gitega and Cibitoke provinces, respectively. Monthly averages were calculated (Table 1). These data were used in the analysis of the seasonal effect on aphid counts during the 12-month-period.
Table 1. Average monthly rainfall and temperature data, calculated with 2007, 2008 and 2009 dataa, corresponding to the period of experiments
aData were obtained from the Institut Géographique du Burundi (IGEBU) stations located at Bujumbura Airport and Gitega Airport, and at the Institut des Sciences Agronomiques du Burundi (ISABU) Mparambo research station. These climate data collection sites are representative of the study sites located in Bujumbura Rural, Gitega and Cibitoke provinces, respectively.
BBTD spread was evaluated by assessing the effect of location, cultivar and type of planting material (i.e. TC plantlets or suckers). The data collected on BBTD symptoms were converted into 1 for presence or 0 for absence. A logistic regression model was then fitted to assess the effect of these variables on disease occurrence:
where P is the probability that the event occurs (i.e. disease presence), 1−P is the probability that the event does not occur, β0 is intercept and β1,β2… are regression coefficients of x1, x2… explanatory variables, respectively.
Monthly aphid counts were analysed using the Poisson regression model which expresses the natural logarithm of the event as a linear function of predictors. The model was loge (y) = β0 + β1Χ1 + β2Χ2…, where y is the mean of the response variable, β0,β1,β2… are the regression coefficients and X1,X2… are the predictors. These models randomly take references for each category of parameters.
Data analyses were carried out using GenStat 12th edition (VSN International Ltd.), while figures were drawn using instat v. 3.36 (SSC).
Effect of altitude on BBTD spread at Benga and Kagazi
Regression analysis indicated that BBTD incidence significantly varied (P < 0·001) according to trial location, banana cultivar and planting material type (i.e. suckers or TC plantlets; Fig. 1). Visible disease symptoms were first observed 3 months after trial establishment at Benga, compared with 4 months at Kagazi. Two years after trial establishment, TC plants of all tested cultivars showed a disease incidence ranging from 35% to 40% at Benga and from 10% to 35% at Kagazi. At 24 months, sucker-derived plants showed a BBTD incidence at Benga ranging from 30% to 50% and from 20% to 50% at Kagazi. All tested genotypes showed disease symptoms, although Nyambururu (AAA-EA) was least infected for both planting material types in Kagazi and for sucker-derived plants at Benga (Fig. 1).
Influence of plot location on BBTD spread and aphid occurrence
During the first 2 months, the plots located inside the infected banana fields attained a disease incidence of about 10%, whereas no BBTD symptoms were reported within the plots located outside the affected banana fields. Nine months after trial establishment, BBTD incidence had increased and was higher in the plots located inside a banana plantation where infected mats had been uprooted (56·4%) than in plots inside affected banana fields where infected mats had not been uprooted (21·8%) (Table 2). In plots established 5 and 30 m north of an affected banana plantation, BBTD incidence was 6·0% and 12·3%, respectively. On the other hand, a disease incidence of 9·9% was observed in plots located 5 m south of an affected banana field. The banana plants established 30 m south of an affected plantation did not show any disease symptoms up to 9 months after trial initiation. Among the Musa cultivars grown, Americani (AAA) showed BBTD incidence ranging from 0% to 15·9%, while Gisandugu (ABB) and Kamaramasenge (AAB) showed disease incidence ranging from 0% to 7·3%, 9 months after trial establishment. The TC-derived Yangambi Km5 (AAA) plants showed a disease incidence ranging from 0% to 19·9%, compared with 0–7% for sucker-derived Yangambi Km5 plants (control). This difference could be explained by the higher susceptibility of young Yangambi Km5 TC plantlets to BBTV infection compared with the more robust sucker-derived Yangambi Km5 plants (Table 2).
Table 2. Banana bunchy top disease (BBTD) incidence observed in different plot locations 9 months after trial establishment
Disease incidence (%) according to plot locationa
aPlot located (i) inside a BBTD-affected field where BBTV-infected mats had not been uprooted, (ii) inside a BBTD-affected field where BBTV-infected mats had been uprooted, (iii) 5 m south of a BBTD-affected field, (iv) 30 m south of a BBTD-affected field, (v) 5 m north of a BBTD-affected field and (vi) 30 m north of a BBTD-affected banana field.
Americani (AAA, Cavendish)
Yangambi Km5 (AAA)
Sucker-derived Yangambi Km5 (AAA)
Total BBTD incidence
Using TAS-ELISA, BBTV was confirmed in 85 out of 225 leaf samples collected, confirming the presence of BBTV in all six plot locations. In addition, 24% of leaf samples collected in plots located 30 m south of an existing infected field, where no BBTD symptoms had been reported, were BBTV-infected. Intense yellow coloration of positive samples, with OD ranging from 0·8 to 2·18, was obtained for samples collected on plants showing marginal leaf chlorosis to bunchy top appearance (i.e. disease severity stages 3 to 5; Niyongere et al., 2011a), while OD values ranging from 0·1 to 0·8 were measured on samples taken from plants showing dark green streaks on their leaf lamina (i.e. disease severity stage 1 or 2).
Regression analysis showed that aphid numbers varied significantly (P < 0·05) with plot location and season in the six treatments (Table 3). In the first month of the trial (April 2009), aphid numbers were highest in plots established inside a banana plantation and north of an existing banana plantation. On the other hand, few aphids were reported during the first 5 months of the trial in plots south of a banana plantation. Aphid counts were also lower in September and October, coinciding with the beginning of the rainy season in Burundi.
Table 3. Aphid counts as influenced by plot location and season during the 9 months of the trial
Antilog of estimate
Plot location (i) and April were randomly taken as references among locations and season (months) factors, respectively.
NS, *, ***: not significant or significant at P ≤ 0·05 and highly significant at P ≤ 0·001 (Fisher’s test), respectively.
Plot location (ii)
Plot location (iii)
Plot location (iv)
Plot location (v)
Plot location (vi)
Seasonal and spatial distribution of banana aphids
Aphids sampled in yellow traps from three different agroecological zones
Banana aphids were observed in all the three study regions in Burundi, irrespective of altitude and the presence of the disease. The highest aphid numbers (average yearly count of 904 winged aphids) were reported at Benga (1268 m a.s.l.), where dense AAA-EA banana plantations are located. Far fewer aphids (average yearly count 256 winged aphids) were observed at Kagazi, which is located 890 m a.s.l. and where Yangambi Km5 is cultivated in a less dense intercropping system. The lowest aphid numbers were reported in the well-maintained Musa collection at IRAZ, Mashitsi (1645 m a.s.l.), with an average yearly count of 234 winged aphids.
Winged aphid numbers fluctuated throughout the year (Fig. 2) within plantations and were significantly (P < 0·05) influenced by the prevailing temperature. The highest winged aphid numbers were recorded in the period July to August, corresponding to the dry season in Burundi (Table 4). This period is characterized by high average temperatures (24–32°C) and the absence of rainfall. In addition, banana management activities come to a standstill during the dry season. Deleafing and desuckering are not practised, leading to bushy banana mats and clusters, creating an ideal breeding ground for aphids. In contrast, deleafing and desuckering are carried out during the rainy season to provide more light for the intercropped legumes.
Table 4. Effects of rainfall and temperature during a 1-year period on average aphid numbers collected in yellow water traps in three different agroecological zones in Burundi
Antilog of estimate
January was randomly taken as a reference amongst months.
NS, *: not significant or significant at P ≤ 0·05 (Fisher’s test), respectively.
Aphids sampled in yellow traps located inside and outside a banana field
The average yearly aphid numbers collected in yellow water traps positioned inside a banana field and at 30 and 50 m south of a banana plantation were 207, 69 and 27, respectively. The number of aphids was inversely proportional to the distance from the banana plantation, with an estimated coefficient of 0·33 at 30 m, compared with 0·13 at 50 m from the banana plantation (Table 5). The aphid numbers varied over a period of a year and significantly lower (P < 0·001) winged aphid numbers were observed at 30 and 50 m from the banana plantation.
Table 5. Influence of distance from a banana plantation and rainfall on average aphid numbers collected in yellow water traps located inside and 30 and 50 m away from a banana plantation at the ISABU Mparambo research station
Antilog of estimate
*, ***: significant and highly significant at P ≤ 0·05 and P ≤ 0·001 (Fisher’s test), respectively.
Traps located at 30 m
Traps located at 50 m
Assessing the effect of altitude on BBTV transmission
BBTV transmission using aphids from non-infected banana fields
A BBTD incidence of 0% was reported at Bujumbura on TC plantlets when aphids collected from the Musa collection at IRAZ were allowed to feed on them. This demonstrated the absence of BBTV in the IRAZ banana field collection. However, virus transmission trials using aphids collected from the IRAZ banana field collection and allowed to feed for 24 h on BBTV-infected plants resulted in BBTD incidence levels of 83% and 50% at Bujumbura and Gisozi, respectively. These aphids were thus able to acquire and transmit the virus independently of altitude, albeit at different rates.
BBTV transmission using aphids from infected banana fields
A BBTD incidence of 100% was observed at Bujumbura on TC plantlets infected with aphids obtained from affected farmers’ fields at Cibitoke. This confirmed that the aphids collected at Cibitoke carried BBTV. The transmission rate after aphids collected from farmers’ fields at Cibitoke were allowed feeding access of 24 h on BBTV-infected plants was 83% and 67% at Bujumbura and at Gisozi, respectively.
BBTD incubation period at Bujumbura and Gisozi
The mean incubation period of the disease was 21 days at Bujumbura (780 m a.s.l., average minimum and maximum temperatures 17–32°C) vs 84 days at Gisozi (2090 m a.s.l., average minimum and maximum temperatures 6–25°C). BBTD symptoms were consistently observed on the second newly emerging leaf. The observed slower growth of banana plants and corresponding reduced leaf emergence rate at higher elevations translates to a longer incubation period. The first observed symptoms on BBTV-infected TC plantlets at Bujumbura, as well as at Gisozi, were dark green dots and dashes of variable length on the leaf veins and dark green streaks on leaf midribs and petioles. This is in line with previous observations made by Magee (1927) and Thomas et al. (1994). More advanced symptoms included leaf marginal chlorosis, dwarfing of leaves and upright and crowded leaves at the apex of the plant, giving the plant a bunchy top appearance (Magee, 1927).
All 30 samples obtained from infected TC plantlets tested positive, while the six samples collected from the control plants tested negative in TAS-ELISA.
In this study, the epidemiology of BBTD at different altitudes, the ability of its vector to acquire and transmit BBTV at low and high altitudes, and the effect of plot location relative to an existing diseased field and wind direction on rates of reinfection were determined.
The higher disease incidence observed at Benga can be attributed to the presence of a large number of winged aphids (904 trapped aphids) compared to Kagazi (256). The larger number of winged aphids observed in Benga could result from the more densely cultivated banana (AAA-EA) production system, compared to the less dense Yangambi Km5-dominated system in Kagazi.
BBTD incidence also varied according to plot location, with a high disease incidence (56·4%) observed in plots located inside an affected banana plantation where infected mats had been uprooted. The observed high disease incidence could be explained by winged aphids being disturbed during the removal of diseased mats and eventually migrating to the newly established young plantlets. Young & Wright (2005) also reported the preference of aphids for newly established young plants. In addition, the removal of diseased mats resulted in an overall lower number of young suckers (i.e. lateral shoots) in the field and a less dense plantation, which may have made it easier for winged aphids to locate the new young plantlets.
A lower disease incidence (21·8%) was observed in plots located inside an infected banana plantation where infected mats had not been uprooted. Winged aphid populations may have been less disturbed in this plot type compared to plots where all the infected mats were uprooted. Additionally, an overall higher number of lateral shoots which also attract winged aphids may have contributed to the lower incidence observed on the newly established plantlets. In addition, aphid spread was found to be influenced by plot position (i.e. north or south of an existing field) and distance between newly established plots and existing affected banana plantations. For example, no banana plants showed BBTD symptoms in plots located 30 m south of an affected banana field 9 months after trial establishment.
In both study areas, susceptibility to the disease varied according to the genotype used, which is in agreement with reports from India and the Philippines (Thomas & Caruana, 2000; Geering, 2009). Moreover, Bhadra & Agarwala (2010) reported that under field conditions, the population of aphid vectors grows in response to the phenology of their respective host plants. Banana pseudostem colour influences aphid preference (Robson et al., 2006), which in turn contributes to variations in infection level.
In terms of seasonal and spatial distribution of the vector, winged aphids were found in all studied regions irrespective of altitude and the presence of BBTD. However, aphid colonies were significantly influenced by climatic conditions. The highest winged aphid numbers were reported during the dry season in Burundi. Cultural practices such as deleafing and desuckering, usually practised during the rainy season (i.e. when legumes are cultivated in association with banana), should preferably be continued during the dry season.
In addition, winged aphid numbers were significantly but negatively influenced by rainfall, which is in agreement with Young & Wright (2005), who stated that in Hawaii, aphid numbers were negatively correlated with rainfall in less dense canopy conditions. It is possible that reduced leaf canopy coverage will allow raindrops and run-off water to negatively affect the aphid populations on the banana pseudostems. In this study, the aphids were observed in traps located up to 50 m from an existing banana plantation. Although Allen (1987) stated that more than 60% of new infections occur within 20 m of the nearest source of infection, the vector could potentially spread BBTD much further from the source of infection.
In the present study, BBTV transmission trials carried out at two different altitudes revealed that aphids collected in healthy fields and subsequently reared on BBTV-infected plants were able to acquire and transmit the virus at Bujumbura (780 m a.s.l.) as well as at Gisozi (2090 m a.s.l.). Thus, although BBTD is currently mainly reported at below 1300 m a.s.l. (Walangululu et al., 2010), the present study indicates the possibility of disease spread to higher altitude regions. The TC plantlets infected at Bujumbura (780 m a.s.l.) showed the first BBTD symptoms after 21 days, while for those infected at Gisozi (2090 m a.s.l.) the mean incubation period was 84 days. The first BBTD symptoms were observed on the second newly emerging leaf after inoculation and a lower leaf emergence rate was observed at the higher altitude site, which may be one of the factors that led to delayed symptom expression. Although the observed virus transmission rate was lower at the high altitude site, climate change (i.e. higher temperatures) and the existence of dense and poorly maintained banana fields in Burundi may enhance the spread of the disease to regions at higher elevation.
The TAS-ELISA tests confirmed BBTV in banana leaf samples and the presence of the virus in symptomless plants. These infected but symptomless plants could serve as a source of inoculum for further virus spread (Allen, 1987). Higher virus concentrations were reported in samples displaying advanced BBTD symptoms. As such, an effective management strategy is dependent on early visual detection and prompt uprooting of plants showing symptoms, as suggested by Magee (1938), Dale (1987) and Hooks et al. (2008).
Under suitable conditions, aphid vector populations rise progressively, resulting in a gradual increase of diseased mats and corresponding inoculum levels (Magee, 1938; Hooks et al., 2009). Once the disease has been introduced into an area, Robson et al. (2006) reported that eradication is very difficult. Hence, the use of virus-free planting material of tolerant cultivars and new plot establishment at least 30 m away from an existing diseased field should be adopted to prevent the early establishment of the virus and possible inoculum build-up. In parallel, regular cultural practices (e.g. prompt uprooting of diseased mats and regular deleafing, desuckering and weeding) need to be enforced in order to keep aphid colonies below the threshold at which winged aphids are produced and spread the virus (Kumar et al., 2011). In addition, quarantine measures should be established between BBTD-affected and unaffected areas (Thresh, 2003). Training of extension staff and farmers is paramount for an integrated management/control strategy under small-scale farmer conditions in Burundi.
The authors would like to thank the Directorate General for Development (DGD, Belgium) for funding this work through the CIALCA (Consortium for Improving Agriculture-based Livelihoods in Central Africa) project.