394 HARPER: FACTORS INFLUENCING DISTRIBUTION 
extends to the inland edge of the Oligocene strata. Croom * 
has reported it from Hartford, Ga. (a settlement, now extinct, 
which was near the Ocmulgee River in Pulaski county, about oppo- 
site the site of the modern town of Hawkinsville), which is about 
the same distance from the Atlantic coast as that last men- 
tioned. 
Between these limits 7. imbricarium is very abundant m 
Georgia ; and I will here give its observed distribution in the state 
somewhat in detail, as it is on these observations that my theory 
of its distribution is chiefly based. 
East of the Ogeechee River it is abundant in the wet pine- 
barrens of Chatham and Effingham counties, and is sparingly rep- 
resented around Millen, in the southern corner of Burke. West 
of the Ogeechee, in Emanuel and the upper part of Bulloch, I 
have seen it only in creeks and small rivers, where, as before 
mentioned, it attains a considerable size. Traveling westward 
along the Georgia and Alabama Ry., I saw it last July at many 
points in the pine barrens of Tattnall, Telfair, Dodge, Wilcox and 
Dooly counties, extending to within eight or ten miles of Cordele. 
Between Cordele and the Flint River the Columbia formation 
seems to be absent, and Zaxrodium imbricarium likewise. Farther 
south it occurs in wet pine-barrens or in small sluggish pine-bar- 
ren streams in Worth, Berrien, Coffee, Ware and Appling coun- 
ties. In the upper part of Coffee county, as for instance in the 
vicinity of Douglas, it grows also in creeks as in Bulloch. 
In Sumter it seems to be confined to a number of shallow 
ponds about three miles from the Flint River in the southeasterm 
corner of the county, where the Lafayette and Columbia forma- 
tions overlap a short distance. Southward from there, it is very 
abundant in Lee county, and occurs in Dougherty, Calhoun, Mit- 
chell, Miller, Early, and the northern parts of Thomas and Dae 
tur, in the terrane of the Lower Oligocene. Throughout this 
part of the state its usual habitat is in shallow pine-barren ponds 
which sometimes dry up in summer, which it occupies t° the 
exclusion of other trees, But in Thomas and Decatur counties 
(east of the Flint River) I have seen it in permanent ponds several 
hundred acres in extent and probably five or six feet deep, con” 
* Am. Jour. Sci. 28: 166. 1835. 
