316 CARNEGIE INSTITUTION OF WASHINGTON. 



roborated the recent conclusion that in the United States at least this climax 

 is confined to the desert regions along or near the lower course of the Colorado 

 River and that the extensive areas of Larrea and Prosopis that reach from 

 Texas to Arizona are actually savannah. The coastal forest and its outposts 

 were followed from central California to northern Washington and the transi- 

 tion association of the same climax through eastern Washington and Oregon, 

 while the structure and zonal relations of the montane and subalpine forests 

 were seen in the Sierra Nevada, on Mount Rainier, and in the Cascade and 

 Blue Mountains of Oregon and Washington. 



Changes in Grassland, by F. E. Clements and E. S. Clements. 



While further evidence of the widespread changes in grassland in conse- 

 quence of climatic cycles, grazing, and fire has been obtained from all of the 

 associations, perhaps the most striking of this has to do with the differentiation 

 of the bunch-grass prairie, the tall-grass savannah of the northern portion 

 of the Great Basin, and the original grassland of Death Valley. An unusual 

 opportunity has also been afforded for the study of natural parks and savan- 

 nah in widely separated regions. The major dominants of the bunch-grass 

 climax differ so completely in southern and central California from those 

 of northeastern Oregon and adjacent Washington as to raise a serious question 

 of their former continuity and actual relationship. In the southern portion 

 of the climax, Stipa setigera and eminens are overwhelmingly predominant, 

 while in the north these do not occur, and Agropyrum spicatum, Festuca ovina, 

 and Poa sandbergi outrank all their associates. While an undoubted differ- 

 entiation has occurred, due in part to climate and in part to mountain bar- 

 riers, a careful scrutiny of hundreds of relict areas, especially in the inter- 

 mediate region, indicates that originally these were essentially a unit. 

 Although Agropyrum spicatum and the two Stipas have nowhere been found 

 in contact to-day, they existed as co-dominants in the intermediate region 

 within the historical period and many of their associates range throughout. 

 Thus, Koeleria, Poa, Sitanion, Elymus, and Stipa comata occur through prac- 

 tically the entire association and Danthonia and Melica through most of it. 

 The divergence of the two portions doubtless owes much also to the fact that 

 fire and overgrazing have opened the way to the annual grasses to become 

 dominant in California and the sagebrush in Oregon and Washington. 



It has been assumed until the present season that the sagebrush savannah 

 that covers the northern and eastern part of the Great Basin had developed 

 from the bunch-grass association. The investigations of the summer, how- 

 ever, make it clear that this is true chiefly for the portion found in eastern 

 Oregon and Washington. The grasses of the major part of this community 

 are the tall-grasses of the mixed prairie, Stipa comata, Agropyrum glaucum, 

 Sporobolus cryptandrus, and Koeleria cristata. The short-grasses, Bouteloua 

 gracilis, Carex filifolia, Bulbilis dactyloides, and Hilaria jamesi, have not 

 been seen west of the Teton and northern Wasatch ranges, though the first 

 two have been recorded for Washington. Thus, it appears probable that 

 both the elements of the mixed prairie were once in contact with the bunch- 

 grass climax in the north as well as in the south, though the short-grasses 

 were clearly fewer and much less abundant. 



The contact between grassland and forest has been traced along a greater 

 front than in any previous year. This has permitted an exceptional oppor- 

 tunity for studying the relation of natural parks and savannah to each. 



