GEOLOGY AND PALEONTOLOGY OF CANAL ZONE


erally faint spiral cords more or less roughened by growth threads. Upper part of peristome gently arched forward.
 Height 1 mm, diameter 1.6 mm (figured bicarinate specimen). Height 1.1 mm, diameter 2.3 mm (figured large specimen).
 Type: Acad. Nat. Sci. Phila. 2831.
 Type locality: Dominican Republic, Miocene.
 This small Cyclostremiscus occurs throughout the Gatun formation. It is rare in the lower part, rare to abundant in the middle part, and rare in the upper part. Several hundred specimens were collected from the middle part at locality 147b.
 These fossils show a considerable range of variation in the outline of the body whorl, in the presence or absence of spiral threads near the upper and lower carinae, and in the strength and coarseness of the spiral cords facing the umbilicus. The usual form is bicarinate (pl. 17, figs. 7-9). Though tricarinate shells (pl. 17, figs. 10-12) are not common in the middle part of the Gatun formation, the few specimens from the lower and upper parts (one and two, respectively) are tricarinate. Spiral threads, generally faint, near the upper and lower carinae are exceptional. The umbilical spirals generally are weak. They are, however, exceptionally strong on the large specimen shown on plate 17, figures 13-15. This large specimen, collected from the middle part of the Gatun formation, shows a further modification in the rounded outline of the body whorl. At the beginning of the body whorl of this specimen the upper carina is moderately strong, the basal carina is weak, and the peripheral angulation is faint. These carinae and angulation rapidly disappear. This specimen is larger than the bicarinate and tricarinate forms. Inasmuch as mature bicarinate and tricarinate forms are rounded near the peristome, the exceptional features of the rounded specimen are presumably correlated with its size. Toula's illustrations, however, show a shell of moderate size, the body whorl of which is rounded at an early stage.
 Cyclostremiscus pentagonus occurs in the Cercado formation of the Dominican Republic and in the Bowden formation of Jamaica. The few available specimens from the Dominican Republic and Jamaica are tricarinate. It was formerly thought that the Gatun form could be differentiated by the weak sculpture on the umbilical wall (Woodrmg, 1928, p. 441). That sculpture, however, is too variable for consistent differentiation.
 Closely related forms are living in the western Atlantic and the eastern Pacific. "Circulus" trilix (Bush) (1897, p. 127, pl. 22, figs. 6, 10, 10a, 12, pl. 23, figs. 10, 15), ranging from Cape Hatteras to Cuba, is consistently tricarinate and enlarges more rapidly, so


that with the same number of whorls (about 4/) Recent shells are almost twice as large. Recent shells and' those from the Gatun formation have the same kind of protoconch and aperture. Inasmuch as the degree of enlargement is the only character now apparent t8 differentiate Recent shells and tricarinate fossils, treat4 ment of "C." trilix as a subspecies of C. pentagonusi appears to be preferable. Fossils from the early' Miocene Chipola formation and the middle Miocene Shoal River formation of Florida have been referred to "Circulus" trilix (Gardner, 1926-47, p. 600, 1947), These Florida fossils are tricarinate and are larger than those from the Gatun formation.
 "Circulus" cerrosensis Bartsch (1907, p. 173, figs. 9a, b, c), which ranges from Santa Catalina Island, Cali-. fornia, to Baja California and the Gulf of California, and probably to Panamd, is the eastern Pacific analog of "Circulus" trilix. It has not been determined, whether Pacific and Atlantic shells can consistently be distinguished. Relatively strong spiral cords facing the umbilicus are more common in the few lots of Pacific shells. Shells from both oceans that are still lustrous show under strong light very faint miicroscopic spiral lineation. It is doubtful whether Cyclo-! stremiscus glyptomphalus Pilsbry and Olsson (1945-52, p. 67, pl. 7, fig. 3, 1952), a Pleistocene form from the Pacific coasts of western Panami, can be distinguished from "Circulus" cerrosensis. Pilsbry and Olsson suggested the probability of local races of a widely spread species. Cyclostremiscus glyptobasis Pilsbry and Olssoq, (1945-52, p. 66, pl. 7, figs. 4, 4a, 1952), also from the Pleistocene of western Panama, probably is a variety, or subspecies, of "C." cerrosensis with a sculptured bas.. The Ecuadorean form "Circulus" cosmius Bartsch (1907, p. 173, figs. 8a, b, c) also is closely allied to' "Circulus" cerrosensis, but is characterized by a slight downward bending of the upper margin of the peristome I where it extends forward. Though the peristome of the type of "Circulus" cosmius is damaged, four speci mens in the type (and only) lot have a perfect peristome. The seven specimens in the type lot have very weak umbilical sculpture. "Circulus" occidentalis Pilsbry and Olsson (1941, p. 48, pl. 9, fig. 3), from the Pliocene of Ecuador, shows downward bending of the peristome and should be compared with "Circulus" cosmius.
 Cyclostremiscus tricarinatus (C. B. Adams) (Pilsbry and Olsson, 1945, p. 271, pl. 28, figs. 3, 3a, 3b), living on the Pacific coast of Panama, is similar to the tri, carinate species so far mentioned. It has, however, faint axial riblets between the periphery and suture, and therefore is intermediate between Ponocyclus and Cyclostremiscus s.s.
 Occurrence: Lower, middle, and upper parts of Gatun formation (middle and late Miocene). Lower-


74


I