Correspondence. B57 
natural bridges of the so-called "Natural Bridge park'" (Chapman's am- 
phitheatre) at the head of Salt creek are not developed in any way more 
strikingly than they are in southern Kansas. Indeed, neither cave nor 
natural bridge was seen that equalled the Comanche cave or the Bear 
creek bridge described in my paper on "The Permian System in Kan- 
sas," though the roof of one tunnol-like cave, to which I was conducted 
by Mr. Chapman, serves as a bridge for the passage of a road. This 
bridge is in the Shimer gypsum, which is the principal cave-foniiing 
member of the formation at that locality. Near the upper limit of the 
brine-spring incline at the head of Salt creek, someof the blocks of gyp- 
sum that have rolled down from the Cave Creek brow of the canyon 
have been curiously fluted by the solvent action of rain. 
In his Red River report, in a letter quoted by Pi-es. Edward Hitch- 
cock on page 160, Captain R. B. Marcy described a great "'gypsum- 
belt" which he stated that he had traced "from the Canadian river, in 
a southwest direction, to near the Rio Grande in New Mexico," men- 
tioning his observation of it upon the Brazos, Colorado, and Pecos riv- 
ers. But on page 99 of the same report, he attributes the belt, on the 
authority of Maj. Long, to the Arkansas valley also, confusing it with 
the gypsum of the Marion (Geuda) formation: and in his "Map of the 
Country between the Frontiers of Arkansas and New Mexico," he 
makes it cross the Canadian in the eastern portion of the great bends 
of that river, which would place it in D county, Oklahoma. Neverthe- 
less, I believe that the gypsum-crowned bluffs represented in his "View 
of Gypsum Bluffs on the Canadian River" (Plate 5) are not those of D 
county, which are capped with the much higher gypsum herein de- 
scribed as the One Horse gypsum, but belong to the Cave Creek belt and 
that by some error, the Canadian river segment of this belt was not 
placed upon the map in accordance with the actual observation of 
Capt. Marcy, who crossed this portion of the belt in 1849, during his 
expedition from Fort Smith to Santa Fe. It is impossible to say 
whether all of the southwestern localities of gypsum which he ascribes 
to this belt are in the extension of the Cave Creek, as some of his ob- 
servations may have related to one of the other southwestern gypsum 
horizons; and it seems more probable that he used the term "gypsum - 
belt" V)roadly for an elongate tract of country in which gypsum was 
common, a use which would imply about the same stratigraphic range 
as is included in the Double Mountain beds of Mr. Cummins. The Up- 
per Red River map, however, shows clearly that the gypsum of Gypsum 
and Groesbeck creeks, in adjoining parts of Greer county, Oklahoma, 
and Hardeman county, Texas, which may be called Quanah yypstun, 
from the town of Quanah, near which it is being utilized in plaster- 
mills, was not included in the main part of Marcy's gypsum-belt, since 
the latter; as represented on this map, is far to the west of gypsum 
creek and Mr. Cummins states that the Permiam of northern Texas 
dips to the northwest, a direction in which the surface of the country 
rises.* Whatever doubt may remain as to the relation of outlying parts 
♦Overlying the Quanah gypsum, are laminated dolomites, which may be called the 
Groesbeck, dolomites, as they are well exposed on Groesbeck creek, in Hardeman 
county. 
