Airborne spores of the fungal pathogens causing Sigatoka diseases in banana and plantain were monitored using rotorod spore traps, sited at various heights within an infected plantation in Costa Rica from December 1993 to February 1994. Different capture patterns of ascospores and conidia were found and the relationship between wind behaviour and spore catches was investigated. This information has enabled an assessment to be made of the reliability of point measurements of airborne spores for monitoring spore movements on the plantation scale. The use of such information in forecasting the airborne movement of these spores and the likely role of the wind in the spread of this disease to uninfected areas is discussed.

Trajectory analysis of the windfields at 10 m and 1.5 km was used to determine the direction and extent of windborne movements of insects trapped in spring and summer in south-eastern and eastern Asia for the period 1968 to 1985. Approximately 2500 trajectories, depicting the movements of airborne particles, were constructed where temperatures were high enough (>=10°C) to allow flight and where wind speeds (> =5 kmh−1) were expected to lead to downwind displacements. Trap catches were dominated (97%) by delphacid species (e.g. Nilaparvata lugens Stål and Sogatella furcifera (Horváth)), with Cicadellidae (1.1%) and Diptera (0.7%) as the next most numerous. Very few Lepidoptera were reported. Most trajectories were less than or equal to 40 h in duration. The previously recognized northwards migration along a broad front in prevailing Summer Monsoon and Trade winds was confirmed, as was the important role of frontal depressions in the windborne migration of insects in temperate areas. Nearly 50% of the trajectories were constructed in mobile weather systems which both disturbed and strengthened the prevailing winds and led to pronounced differences in the direction and distance of sources for insects caught at any catching site. Similarly, fluctuations in position of quasi-stationary fronts induced considerable mixing of populations from different sources and influenced the northern extent of each phase of rice pest migration. The results supported earlier findings that migrations occurred between the tropics and temperate areas during the spring and summer, but also indicated that migrations within the tropics at this time of year were a more frequent occurrence than had been considered previously.

Banana leaves showing different levels of black Sigatoka disease were collected from an unsprayed plantation in Costa Rica during two separate periods representing the wet to dry season transition (October 1993 -February 1994 and the dry to wet season transition (April -September 1995). Laboratory studies were used to investigate the relationship between the release of Mycosphaerella fijiensis ascospores and the amount of inoculum on banana plants showing different levels of infection, as assessed by leaf necrotic area. The number of perithecia present in the necrotic area was used as an indication of potential ascospore loads and was investigated as a series of regression equations. A series of rewetting and incubation regimes was used to investigate spore release under field conditions (21°C and 100% relative humidity in the early morning and 28"C, 60% relative humidity on days when it rained in mid-afternoon). Results suggest that rainfall, combined with a high temperature, may lead to peaks of ascospore release but without necessarily increasing overall numbers released over periods of up to 4 days and that a high level of spore release was less sensitive to changes in temperature once it had been initiated. The exact role of temperature in spore release is still unclear, however, as leaf samples kept at atypically low temperatures also released non-germinating ascospores. An average of 4.5 ascospores was released per perithecium. This does not resolve ambiguities in the literature regarding the number of ascospores present in each perithecium. A linear model relating the average ascospore numbers and necrotic area, using quick estimates of the amounts of necrotic area on the leaves of a random sample of plants across a plantation, is proposed, to give an indication of the relative amount of airborne inoculum potentially available between different plantations.