Experiments were carried out at Malanville in 1996 to examine the effect of predators on healthy and infected grasshoppers. There was an approximately 40% increase in the risk of predation to infected compared with healthy hoppers. (Thomas et al 1998). This result was examined further with brown locust in the Karoo in 1998. These studies showed similarly that infection increases risk of predation and cannibalism under field conditions. The extent of this will depend on the nature and abundance of the natural enemy fauna at particular sites but where activity is high, it is likely to play an important role in both mediating speed of kill following application and influencing the dynamics of secondary cycling (see below).

Data collected during the 1997 trials in Niger indicated that the vast majority of infected individuals are predated or scavenged around the time of death and only 1 - 2% of the grasshoppers killed by Metarhizium go on to produce new sources of inoculum under these field conditions. This suggests that the chances of secondary cycling are greatly reduced through the interaction with predators and scavengers. However, this is balanced to some extent by the fact that if cadavers are fully mycosed and/or sporulating, they are avoided by scavengers and persist for far longer than uninfected cadavers (i.e. they persist for many days whereas uninfected cadavers are removed within 24-48hrs). Thus, even though infected cadaver densities tend to be low, there are still opportunities for pathogen cycling. This was confirmed with a further study in Niger in which infected cadavers where laid out in a grid at a realistic field density of 0.25 per m² in an untreated 40x40m plot. Samples of grasshopper were then collected from this plot after 48 hrs and incubated for 21 days to record infections. It was found that even though this was an open field plot and some of the grasshoppers may have been exposed to the 'cadaver graveyard' for only a limited time, 20% of the grasshoppers went on to show mycosis compared with 0% in equivalent control samples collected away from the graveyard. Thus, under Sahelian conditions, it appears that limited but, in population dynamic terms, important secondary cycling is likely to occur, increasing the impact of single spray applications.

Similar experiments conducted in South Africa also suggest that infected cadavers are avoided by scavengers and predators. Here, however, it appears that extreme hot dry conditions in the Karoo act to limit sporulation of these cadavers. That is, although infected cadavers persist, they do not tend to produce spores unless it actually rains within a few days after death. This suggests that although not impossible, secondary cycling and long-term persistence of the pathogen in the Karoo are very unlikely (S. Arthurs, 1998 (7)).

During the Maine Soroa field trial in 1997 (Figure 5), spore persistence was recorded for >6 weeks. Although this may be due to prolonged spore survival, spore recycling cannot be ruled out. A definitive experiment to distinguish prolonged secondary pickup from true recycling has now been conceptualised, and will be conducted at the next opportunity.

Experiments conducted in laboratory and field on Acrotylus sp., H. daganensis, O. senegalensis, L. pardalina and Z. variegatus show that feeding is reduced significantly during the disease incubation period. This results in an increase in effective control well before disease induced mortality causes reductions in population density. (Thomas et al. 1997, Blanford et al., 1998; S. Arthurs (7)).

Data on thermoregulation by grasshoppers has been published for Oedaleus senegalensis and Zonocerus variegatus (Blanford et al., 1997 A &B), and indicates the ability of these species to raise their body temperatures sufficiently to slow down, but not stop totally, the development of fungus in the insect's body.

Thermal gradient experiments confirmed these data, and these data have been used to define more realistic bioassay conditions, and will be incorporated into a GIS model to predict expected speed of kill of the mycoinsecticide against different target hosts in different ecological zones in Africa. Thermoregulation (behavioural fever) explains most of the difference between field and cage mortality times.

In addition to effects on feeding and interactions with thermal behaviour, laboratory studies on Desert locust and field studies on brown locust indicate that 'sub-lethal' infections (i.e. low doses that may take a number of weeks to kill infected adults under harsh environmental conditions) may also have a negative impact on reproductive fitness. These studies show that size of fat bodies (important for flight dispersal and egg production), fecundity and viability of eggs may be reduced due to infection (unpublished data). These studies are still on going but have important implications for the effectiveness of the pathogen against adult locusts and indicate further that the impact of the disease can be complex and is not restricted to simple effects on mortality.

Observations also indicate an interaction with the hormonal system of locusts.

A user-friendly version of the population dynamic models developed in collaboration with Simon Wood of St. Andrews University is now available in a draft version, and is being developed as an education and training tool. New data on cadaver loss is being incorporated, and more work is needed to incorporate differences between species, but it is hoped to have a reasonable working model in the new year. The aim is to use the model to inform users of how the mycopesticide and pathogen behave and to allow them to explore the benefits of biocontrol relative to chemical alternatives.

A descriptive models incorporating secondary pickup, decay and immigration give close fits to data for trials on Zonocerus (Langewald et al., 1997) and Oedaleus (Langewald et al., SUBMITTED).

Dr. Dan Johnson is working on the incorporation of disease progression data into a GIS model to detail the impact of meteorological variables. Current constraint is the lack of reliable and consistent meteorological data.