We only have superficial knowledge of how much of the natural diversity of the South Texas shrublands might affect the stability of production by any given component such as grasses. One relevant but subtle interaction involves the regulation of the shrub density by relatively inconspicuous host specialist insect herbivores. It is possible to realize the extent of this influence in the natural system only when we introduce plants to other continents free of their specialist parasites. Thus, for example, when prickly pear was introduced to Australia, it spread rapidly across the range and achieved densities unknown in the New World. Although numerous insects were tested (Mann, 1969 and references therein), only one, an inconspicuous moth, Cactoblastis cactorum, proved capable of reducing Opuntia to densities comparable to those in its native habitats. Specialist insects such as moths have also been implicated in regulating the density of Opuntia in the Great Plains of the United States (Cook, 1942) and Phoradendron on mesquite in South Texas (Whittaker, 1982). In general, there is no sound understanding of why certain shrubs are seen at particular levels of abundance and why so few increase densities to the degree occasionally reached by mesquite.

In unmanipulated native habitat in the Rio Grande Plain, a stable balance between shrub clumps and open grassland appears to exist. Except on special soil types or along streams, no one species totally dominates the system. This may be partly due to the fact that the insects that control the density of one plant are encouraged by another plant. Thus Condalia provides nectar for adult moths and butterflies whose eggs will become larvae consuming Celtis or other shrub species. We do not know whether removing floral resources for adults of such insects will lessen their impact on larval host plants in the South Texas system. However, such interactions are to be expected for phytophagous insects with significant food requirements as adults. The point is that indiscriminate removal or disruption of the diverse woody vegetation on the Rio Grande Plain may have the consequence of creating an uncontrolled increase in density by a normally "well regulated" species. Such relationships may help explain why mesquite tends to be more dense where past attempts to remove all woody vegetation have been made. Are checks and balances on mesquite somehow connected to other woody species? We may lose the structure of the native system (if not many component species) before this question is adequately explored.