118 
between it and certain soils and other en- 
vironmental factors can now be made. Its 
hills are almost always in dry sandy loam, 
presumably of Pliocene or Pleistocene age, 
It, avoids on the one hand the fertile lime- 
stone and clay soils which characterize some 
parts of the coastal plain, and on the other 
the hopelessly sterile sands of the shifting 
dunes along the coast and the “scrub” of the 
Florida peninsula. It is most abundant in 
regions where according to the statistical 
maps in the fifth and sixth volumes of the 
Tenth Census less than one acre to the square 
mile was cultivated in cotton in 1880; but in 
soils which at the present stage of the eco- 
nomic development of the southeastern 
United States are being appropriated by 
farmers most rapidly. Being confined to dry 
soil, it is absent from land which is too damp 
for cultivation by ordinary methods (as well 
as: that which is too sterile or too rocky). 
But I have never heard any complaints about 
its interfering with agricultural operations. 
The southeastern salamander seems to be 
invariably associated with the long-leaf pine 
(Pinus. palustris), and it may derive part of 
its. food from the roots of that useful tree. 
The only known station in the Piedmont re- 
gion of: Georgia, mentioned above, is prob- 
ably not right in the city of Chipley, but very 
likely on the Pine Mountains near by, where 
the rocks and soil are pretty sandy, and long- 
leaf pine abounds.’ The range of our animal 
is by no means coextensive with that of the 
long-leaf pine, though, for the tree ranges 
from Virginia to Texas, as well as consider- 
ably farther inland in Georgia and Alabama 
and a little farther south in Florida than the 
salamander does. Two other trees usually 
found in the vicinity of salamander hills, and 
® At the northernmost Alabama stations there is 
a considerable admixture of gravel in the soil, and 
it is possible that if specimens could be obtained 
from these somewhat isolated localities they might 
be found to differ perceptibly in some characters 
from the only form at present known in that state, 
G. Tuza Mobilensis. 
°See Bull. Torrey Bot. Club, 36, pp. 585-586, 
1909. 
SCIENCE 
[N.S. Vou. XXXV. No. 890 
having more nearly the same range, are two 
scrubby oaks, Quercus Catesbaei and Q. cin- 
erea. Mr. Bangs (op. cit., p. 180) states that 
Cumberland Island, Georgia, is the only one 
of the sea-islands (which fringe the coast 
from about Charleston to Jacksonville) on 
which a Geomys occurs. It is also, to the 
best of my knowledge and belief, the only one 
which has Pinus palustris and Quercus Cates- 
bei on it; and its geological history must 
have been somewhat different from that of 
the others. 
Lastly there are some interesting relations 
between the salamander and forest fires, as 
was noticed briefly long ago in the papers 
cited in the first footnote.° Every long-leaf 
pine forest, without exception, is periodically 
swept by fire, which burns off the dead herb- 
age and keeps down the underbrush, but does 
no harm to sound pine trees after they get 
beyond a certain age. (In prehistoric times 
these fires, presumably set by lightning, prob- 
ably did not visit any one spot oftener than 
once in four or five years, on the average; but 
now so many fires are set accidentally or pur- 
posely by man that few of these forests escape 
fire longer than a year or two at atime.)” The 
dunes and scrub of Florida, mentioned above, 
as well as the other extreme, the rich ham- 
mocks, have so little herbage that fires are 
very rare, and in such places there are neither 
long-leaf pines nor salamanders. 
Fires in the southern pine’ woods are most 
frequent in late winter and early spring, and 
the salamanders seem to be most active just 
about that time. The locality near Lock 14 
on the Warrior River, when first discovered 
on April 15, 1911, had evidently been burned 
over a few days or weeks previously, and the 
salamander hills there looked pretty fresh. 
What is still more interesting, none could be 
found in precisely similar areas near by 
10The same relation was uoticed still earlier by 
Sir Charles Lyell, the English geologist, in Screven 
County, Georgia, in the winter of 1841-42. 
(Travels in North America,’’ Vol. I, p. 161, 
1845.) 
Jn this connection, see Bull. Torrey Bot. Club, 
38, p. 522, 1911. 
