Gaining a better understanding of the spatial population structure of infectious agents is increasingly recognized as being key to their more effective mapping and to improving knowledge of their overall population dynamics and control. Here, we investigate the spatial structure of bancroftian filariasis distribution using geostatistical methods in an endemic region in Southern India. Analysis of a parasite antigenemia prevalence dataset assembled by sampling 79 villages selected using a World Health Organization (WHO) proposed 25 x 25 km grid sampling procedure in a 225 x 225 km area within this region was compared with that of a corresponding microfilaraemia prevalence dataset assembled by sampling 119 randomly selected villages from a smaller subregion located within the main study area. A major finding from the analysis was that once large-scale spatial trends were removed, the antigenemia data did not show evidence for the existence of any small-scale dependency at the study sampling interval of 25 km. By contrast, analysis of the randomly sampled microfilaraemia data indicated strong spatial contagion in prevalence up to a distance of approximately 6.6 kms, suggesting the likely existence of small spatial patches or foci of transmission in the study area occurring below the sampling scale used for sampling the antigenemia data. While this could indicate differences in parasite spatial population dynamics based on antigenemia versus microfilaraemia data, the result may also suggest that the WHO recommended 25 x 25 km sampling grid for rapid filariasis mapping could have been too coarse a scale to capture and describe the likely local variation in filariasis infection in this endemic location and highlights the need for caution when applying uniform sampling schemes in diverse endemic regions for investigating the spatial pattern of this parasitic infection. The present results, on the other hand, imply that both small-scale spatial processes and large-scale factors may characterize the observed distribution of filariasis in the study region. Our preliminary analysis of a mountain range associated large-scale trend in the antigenemia data suggested that a nonlinear relationship of infection prevalence with elevation might be a factor behind such observed global spatial patterns. We conclude that geostatistic methods can provide a powerful framework for carrying out the empirical investigation and analysis of parasite spatial population structure. This study shows that their successful application, however, will crucially depend on our gaining a more thorough understanding of the appropriate geographic scales at which spatial studies should be carried out.