Subsistence farmers among the Central American Indians try to grow the maximum number of different varieties of plantains and bananas as a form of in- surance against possible disaster. They initially secure most of the vegetative shoots used to reproduce these Musas from their parents and relatives. However, shoots are given willingly and freely to anyone who asks for them, and new forms disperse readily within a cultural group. By the time an Indian is thirty years old he has usually ob- tained a normal diversity of forms. Some variables in observations on the processes involved in crop improvement are obviated when the subjects of the studies are ba- nanas and plantains. The processes that are made more easy to investigate include perception, selection, maintenance, and dispersal of the varieties. OBSERVATIONS of the dispersal of plants reproducible by vegetative means allow some insights into the domestication process that are more clearly identifiable than is the case with species reproduced by seed. Bananas and plantains are morphologically stable and clearly recognizable, and thus the confusion resulting from genetic variation of the dis- semules is reduced during the developmental stages of this amplification of the theory of domestication. In broadest outline it is possible to consider that domestication is an ongoing process of plant and animal change under human control toward some preconceived set of character- istics.2 For example, people frequently want larger and sweeter fruit or more of whatever is perceived as useful in the plant, and they search continually, even when they have little success in improvement. The component parts