Sponges, simple and homogeneous relative to other animals, are particularly adept at regeneration. Although regeneration may appear to be obviously beneficial, and many specific advantages to regeneration of lost portions have been demonstrated, comparisons of regeneration among species of sponges have consistently revealed substantial differences in style (i.e., relative rates of reconstituting surface features, infilling depressions, regaining lost primary substratum), and overall time course, raising questions about adaptive significance of variations in patterns of regeneration. Do sponges simply regenerate as quickly as possible, given constraints imposed by skeletal construction, morphology, or other traits that are determined primarily by evolutionary heritage? Does allocation of energy or materials impose trade-offs between regeneration versus competing processes such as growth or reproduction? Is regeneration time-course and style an integral part of coherent life history and morphological strategies? One approach to answering these questions is to compare regeneration among species that represent a spectrum of higher taxa within the demosponges as well as different growth forms and life-history strategies. Because detailed ecological studies of sponges have tended to focus on small sets of species of the same growth form, community-wide comparisons have been hampered. Data on growth rate, colonization, mortality, susceptibility to predation, and competitive ability have recently been accumulated for species of sponges typical of the Caribbean mangrove prop-root community. Experimentally generated wounds in individuals of 13 of these species allow comparison of the timing and style of regeneration among sponge species that span a range of life histories and growth forms. The species chosen represent four orders of the class Demospongiae, and include four sets of congeneric species, allowing distinction of patterns related to life history and morphology from those determined by shared evolutionary heritage.