||TheRed Seacoastal plains ofAfricaand theArabian Peninsulaare an important breeding area for desert locust,Schistocercagregaria . This area has been implicated as a source or transit area for locust swarms that threaten agriculture. The spatial distribution of the desert locust on the southern part of theRed Seacoastal plain of Sudan, betweenPort SudanandTokar, was investigated to determine habitat associations of the desert locust and collect information that might help in planning survey and control operations. Observations were made during three subsequent rainy seasons, from November till March, in 1999/2000, 2000/2001 and 2001/2002.First of all, the main plant communities in the study area, a 120 km stretch of coastal plain, were classified and mapped. Sample sites were laid out in an approximately square 5 × 5 km grid of sample sites to collect vegetation data. Multivariate data analysis resulted in the delineation of four main plant communities, which were named after characteristic plant species: theSuaedamonoica scrubland near the coast, the Acaciatortilis scrubland near the Red Sea Hills, thePanicumturgidum grassland at intermediate location and altitude, andHeliotropium/millet, small pockets of cropland (mostly planted with millet) at the transition between thePanicumgrassland and theSuaedascrub. The croplands were characterized by relatively good moisture provision due to run-on water from spreadingwadis, and high abundance and vegetation cover of the locust host plantHeliotropiumarbainense .Rainfall was markedly different between the three seasons: 158 mm in 1999/2000, 138 mm in 2000/2001 and 30 mm in 2001/2002. This resulted in substantial differences in observed locust densities. During the first and wettest season,solitariousadult locusts were observed in substantial numbers, up to 100 per ha, in theHeliotropium/ millet plant community, but hardly in the plant communities characterized by S.monoica , P.turgidum , or A.tortilis . Egg laying and hopper development were found in theHeliotropium/millet plant community but not in the other communities.Gregarizinghoppers were found in one of the croplands. In the second year of the study,solitariousadult desert locusts were observed at low densities in theHeliotropium/millet plant community, and none were found in the other three communities. No egg laying or hopper development was observed in any community in the second season. In the third and driest season of the study, no adult locusts or hoppers were observed in the main study area. However, during the third season, it became possible to take samples in the delta of thewadiBaraka, nearTokar, to the south of the study area. This area receives more water than the main study area and adult solitary locusts were present here. They were associated with millet agriculture, as in the main study area.Plant samples were taken in the central part of the main study area to compare nitrogen content of host plants in the two most suitable locust habitats: theHeliotropium/millet community and thePanicum grassland. Results showed a consistent difference in N content between host plants in the two plant communities, with the nitrogen content being approximately one percentage point higher in theHeliotropium/millet community than in thePanicum grassland. The difference in N-content is probably ecologically significant, as was demonstrated in a life table study in the laboratory, which compared survival, development and reproduction among gregarious locusts fed on millet leaves with 3.9% or 1.4% N. Net reproduction (65 versus 20 female offspring/female), intrinsic rate of increase (0.072 versus 0.047 day-1) and generation time (58 versus 63 days) all indicate greater potential for population increase when nitrogen content in the food is high than when it is low.Geostatisticalmethods were used to investigate whether it is possible to make predictions of locust densities at distant locations by spatial interpolation usingkriging. Locust density maps were constructed, estimation errors were quantified and the relationship between sampling intensity and estimation error was established for the collected data sets.This study shows a strong relationship between desert locust and the millet croplands on the coastal plain of Sudan. This habitat is characterized by better moisture status of the soil and higher nitrogen content in host plants, and contained the far majority ofsolitariouslocusts in the area. All of the observed locust breeding was in millet croplands. Observations on other parts of the coastal plain confirmed the results. The millet croplands cover a very restricted portion of the coastal plain (only 5% of the area), and are relatively easy to survey.