230 JOURNAL OF THE ARNOLD ARBORETUM vol. it 



the processes of elevation and erosion had been f^oing steadily forward 

 in the Ozark region, and conditions had evidently become nuich more 

 arid. The ligneous and herbaceous flora of the northeast, which perhaps 

 had never succeeded in fully occupying the rocky uplands, was now 

 driven back to the protection of the deeper valleys and bluffs. This 

 process probably continued until most of the Ozark region was divested 

 of forest growth, and upon the arid rocky uplands appeared a flora con- 

 sisting principally of grasses and other herbaceous plants, centering and 

 probably originating in the great plains, occupying the then recently 

 emerged lands to the west. Judging from the flora, remnants of which 

 have been preserved in the southwestern barrens and bald knobs, and its 

 afhuity to that of the western plains, the Ozark region must at this i)eriod 

 have been a semi-arid, wind swept, treeless plain; and the great deposits 

 of loess just beyond its northwestern border may have been contempora- 

 neous with it. 



When climatic conditions again became favorable, responding to com- 

 plex geographic causes, tlie southern forest, which liiul been pushed back 

 many hundreds of miles by the previous cycle, once more began to ad- 

 vance nortliward. Doubtless this forest had undergone great changes 

 in its composition since its precursor had occupied the lowlands of the 

 unevolved Ozark region in Tertiary times; many of its ancient forms had 

 become extinct and new ones developed, and nearly all had undergone 

 more or less modification; and yet it was essentially a very similar forest 

 in gener al aspect and floristic affinities. Evidence of this is not wanting 

 both in the field of paleobotany and in the teslimony of living forms. 

 Most of the families and genera of the present forest are represented in 

 the fossil floras preserved in the Tertiary shales of the Coastal plain, 



many of them closely related to those now living and growing above 

 them. 



To mention a single instance of an ancient type, evidently once widely 

 distributed, and of which only a few scattered remnants now remain 

 attention may be called to the Cork-wood {Lcitneria floridana). Very 

 restricted colonies of this curious plant have been found along the Gulf 

 coast from Florida to Texas, and isolated in the central IVIississippi valley 

 in northeastern Arkansas and southeastern Missouri. Its occurrence in 

 the last named locality is one of the most remarkable examples of jiersist- 

 ence of a forest type in our North American flora, and is significant as 

 indicating the antiquity of the swamj) flora of that section. 



At tlie time when the present advance of the southern forest began 

 much of the lowlands bordering the Ozark region on the south was covered 

 with a prairie flora, different in character from that of the rocky plateau 

 and comparable to that now occupying the more recently emerged lands 

 along the Gulf coast, upon which the ligneous plants are gradually en- 

 croaching. This i)rairie phase must have persisted for a considerable 

 time, offering an obstacle to the advance of the forest, but gradually the 

 latter over-ran it until its advance guards stood at the foot of the rocky 



