JANUARY 19, 1912] 
peninsular Florida. I have seen them in 
every county in that state except Franklin, 
Manatee, Lee, Osceola, Brevard, St. Lucie, 
Palm Beach, Dade and Monroe, all of which 
were included by Dr. Eugene A. Smith, in 
his classical paper on the geography of Flor- 
ida,” in what he called the “ pitch-pine, tree- 
less and alluvial region.” These counties are 
all in south Florida with the exception of 
Franklin, which is in middle and west Flor- 
ida, over 200 miles from the rest. Dr. Smith 
admitted that this was not a very homogene- 
ous area, but he grouped these counties to- 
gether for convenience because they produced 
almost no cotton (only two bales being re- 
ported from that whole area in 1880). The 
distribution of Geomys now furnishes an 
additional character for distinguishing these 
counties from the remainder of the state; 
for I have seen salamander hills in every 
county included by Dr. Smith in his other 
two regions, namely, the “long-leaf pine re- 
gion,” and the “oak, hickory and pine up- 
land region.” (The counties in Florida are 
of course more numerous now than they were 
at the time of the Tenth Census, but it so 
happens that that does not affect the truth of 
this statement.) 
About a year ago, in a report on the peat 
deposits of Florida,” the writer divided the 
state provisionally into fourteen geographical 
divisions. Of these the lime-sink region and 
the lake region, near the axis of the peninsula, 
seem to be the headquarters of the sala- 
mander. The animal is not known at all in 
the East Coast strip or the coast region of 
West Florida, both of which consist chiefly of 
modern (active) dunes next to the ocean and 
ancient (fixed) dunes a little farther back. 
It occurs in varying degrees of abundance 
in the remaining divisions, except those south 
of the latitude of Lake Okeechobee. 
* Proc. Bost. Soc. Nat. Hist., 28, p. 176, 1898. 
5 Tenth Census U. S., 6, pp. 175-257, 1884. This 
monograph, like those on other southeastern states 
in the same volume, bears the entirely too modest 
designation of a report on cotton production. 
Ann. Rep. Fla. Geol. Surv., 3, pp. 201-375, 
pls. 16-28, January, 1911. 
SCIENCE 
eT, 
The range of the southeastern salamander" 
seems to terminate abruptly on the east at the 
Savannah River, and on the west at the War- 
rior and the streams which under two or three 
other names connect that river with the Gulf 
of Mexico. From all accounts it appears that 
the various subspecies (or species aS some re- 
gard them) of this group, which are the only 
representatives of the genus east of the Wa- 
bash River and its continuations, nowhere 
occupy the same territory, but are separated 
by rivers, which must be almost impassable 
barriers to an animal which spends its life 
underground and has no use for water. In- 
deed it is difficult to imagine how such an 
animal could ever have crossed any of the 
large rivers which extend all the way across 
the coastal plain; but at some time in the 
past the ancestors of the present salamanders 
must have crossed at least three such rivers, 
namely, the Altamaha, Apalachicola and Ala- 
bama. (The crossing of the Mississippi and 
Ohio rivers by members of the genus must 
have taken place at a much more remote 
period, judging from the much greater geo- 
graphical and phylogenetic gap between the 
species on opposite sides of these rivers.) 
Mr. Bangs, in his paper already cited, ex- 
pressed the opinion that G. Floridanus is 
separated from the other forms by the St. 
Mary’s and Suwannee rivers, with Okefinokee 
Swamp connecting their headwaters. If this 
is.true then the salamanders of middle Flor- 
ida (4. e., that part of the state between the 
Suwannee and Apalachicola rivers) must be 
typical G. Tuza, the same as in Georgia. 
However that may be, salamander hills of ex- 
actly identical appearance can be found within 
a mile or two of each other on opposite sides 
of the Suwannee, which has almost no swamps 
where it passes through the lime-sink region. 
So much for the areal distribution of our 
salamander. Some interesting correlations 
"According to Dr. Merriam’s monograph pre- 
viously cited, Geomys Tuza and its near relatives 
are confined to the three states already named, but 
the same common name is also applied to another 
species which inhabits Arkansas and Louisiana. 
