igO STEBBINS 



The evolution of Bromiis shows how successive cycles of polyploidy can 

 build up high chromosome numbers and can elaborate a network of evolution- 

 ary relationships by bringing together genes derived from widely divergent 

 evolutionary lines. Less complete studies of other genera and groups of 

 genera, particularly the blue grasses {Poa), beard grasses (Andropogon), and 

 sugar cane {Saccharum and related genera), indicate that these groups have 

 had a similar evolutionary history. 



Conclusion. The picture of grass evolution which emerges from these 

 studies can be summarized about as follows. Early in the history of the family, 

 the major groupings, or tribes, became separated from each other partly 

 through adaptation to different climatic zones and partly through the estab- 

 lishment of subtle differences in anatomy and in the structure of the reproduc- 

 tive parts, the adaptive significance of which is not obvious. The later evolu- 

 tion of most of the tribes consisted largely of reduction and simplification of 

 the parts of the inflorescence and was marked by parallel trends in several 

 different evolutionary lines. In addition, many lines developed highly special- 

 ized fruits, of which the inflorescence scales, or bracts (glumes and lemmas), 

 became variously modified to assist in seed dispersal. These major trends were 

 accompanied by a series of cycles of divergent and convergent or reticulate 

 evolution, the latter being brought about by hybridization and chromosome 

 doubling, or amphiploidy. 



The high frequency of amphiploidy in the grass family can be attributed 

 to several causes. In the first place, different species of grasses grow together 

 in large numbers, giving ample opportunities for hybridization. Second, they 

 have light, wind-borne pollen, which can be carried long distances, and many 

 species are self-incompatible, so that they must be fertilized by pollen from 

 another individual of a different genetic type. Third, the individuals of many 

 species are very long-lived and have efficient methods of vegetative reproduc- 

 tion. This means that interspecific hybrids and newly originated polyploids, 

 many of which are highly sterile, can persist in spite of this reduced fertility 

 and produce many offspring, some of which will have an increased fertility. 

 Grasses can therefore pass relatively easily through the "bottlenecks" of par- 

 tial sterility which necessarily accompany evolution by polyploidy. Finally, 

 most grasses are by their very nature adapted to a role as pioneer colonizers 

 of newly available ecological niches. The processes of hybridization and poly- 

 ploidy, acting together, aid them greatly in this role. 



The research of the past twenty-five years has, therefore, given us a rea- 

 sonably clear picture of the evolutionary forces which have molded the grass 

 family into its present form. Plant breeders are now making use of this in- 

 formation to evolve new types of grasses and cereals better suited to human 

 needs. The progress of such research is inevitably slow, since even if we are 

 able to speed up the natural processes of evolution a hundred times, the 

 production of new, valuable types will require scores of years. Nevertheless, 



