HARVARD UNIVERSITY Library of the Museum of Comparative Zoology Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. XCII, No. 1 THE LOWER MIOCENE MAMMAL FAUNA OF FLORIDA By Theodore E. White With Fourteen Plates CAMBRIDGE, MASS., U. S. A. PRINTED FOR THE MUSEUM November, 1942 tttn o* Cooj^^ i**^ Zoology NOtf 20 3842 UR*' No. 1. — The Lower Miocene Mammal Fauna of Florida By Theodore E. White TABLE OF CONTEXTS Palaeontology a. Preface . b. Systematic list c. Summary Geology . a. Local details b. Paleogeography c. Environment d. Conclusions . e. Geological literature page 3 3 4 27 29 29 32 42 44 48 1. PALAEONTOLOGY PREFACE In the Spring of 1931, Mr. Clarence Simpson of the Florida State Geological Survey discovered some fragments of bone on the dump of an abandoned well on the Raeford Thomas farm, 8 miles north of Bell in Gilchrist Co., Florida. Later in the same year he opened a pit about 60 feet west of the old well. From this pit he obtained some vertebrate fossils which were studied by Dr. G. G. Simpson of the American Mu- seum of Natural History. Pure science had to give way to economic research in the Florida Survey, however, and the work in this area was set aside for an indefinite period. In 1938, Dr. Thomas Barbour, who had spotted the specimens in the Museum of the Florida Survey in Tallahassee, was given permission to continue the excavation. The next year, accompanied by Mr. and Mrs. William E. Schevill, he found and reopened the site and obtained some additional material. The Museum of Comparative Zoology has worked at this locality each year since, and plans to do so as long as the returns justify the expendi- tures. From the beginning the Museum has enjoyed the cordial co- operation of the Florida State Geological Survey. 4 bulletin: museum of comparative zoology I would like to call attention to Dr. Thomas Barbour's contribution to science in this enterprise. He initiated the reopening of the quarry and has been an enthusiastic supporter from the beginning. Credit is due, not so much that he gambled and won, but that he continued to gamble, in the face of adverse returns, on a project in which he had faith, for the results of the first year's work were not reassuring, nor were those of the second. It was not till the third year that the richness of the deposit and the scientific value of the fauna became apparent. In 1940 Dr. Barbour bought forty acres of land about the location of the excavation on the Thomas Farm from the Georgia Loan and Trust Co. in Macon, Georgia, which years ago had foreclosed a mortgage on the property. This land has now been deeded to the University of Florida at Gainesville under certain stipulations concerning its use by representatives of Harvard University, the University of Florida, or the State Geological Survey. Acknowledgment is also due to the Com- mittee on the Milton Fund of Harvard University as well as to those of the Marsh Fund and the Bache Fund of the National Academy of Sciences for grants which have cared for part of the expenses in connec- tion with the excavations and the preparation of the material. Many helpful suggestions on the local geology and the paleogeo- graphy have been given by Prof. P. E. Raymond, Mr. H. C. Stetson, and Mr. Frank Whitmore. The intelligent observations of specimens during preparation have made Mr. Russell Olsen's contribution very much more than an exhibi- tion of manual skill. The superior quality of the graphic art of Mr. Eugene N. Fischer has portrayed the characters of the specimens much better than many printed pages. The citations to literature relative to the fauna are given in the text. That prior to 1935 is given by year and letter corresponding to that in the published bibliographies. For material written since that date full reference is given. SYSTEMATIC LIST CARNIVORA The Carnivora range in size from a small mustelid, no larger than a weasel, to the huge Amphicyon, as large as a grizzly. To date six genera, embracing eight species, have been identified. At least half as many more species are represented by material which will not permit even generic identification. white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA o MUSTELIDAE Mephititaxus ancipidens White Proc. New Eng. Zool. Club, 18, p. 92, PI. 14, figs. 1 and 2, 1942. Unfortunately no additional material referable to this species has been found. In addition to this form, four genera of Mustelidae are known from fragmentary material. Two of the genera are represented by lower jaws without teeth and two by lower carnassials which cannot fit either of the jaws. Also there are three isolated first upper molars which may or may not represent one of the above genera. CANIDAE Daphaenus caroniavorus spec. nov. Type. M.C.Z. 3727 (Plate 1, fig. 1), left M1"3. Horizon and Locality. Lower Miocene, L. Arikareean; Thomas Farm, Gilchrist Co., Florida. Diagnosis. About the size of D. hartshornianus, protocone of M1 less well developed, paracone and metacone conical and larger, medial portion of tooth broader, protocone and metacone of M2 vestigial, paracone larger, medial end of tooth as broad as lateral, M3 button-like. This species is only provisionally referred to this genus. The teeth appear to be somewhat degenerate, and when it becomes better known it may be necessary to erect a new genus for it. Paradaphaenus nobilis (Simpson) Plate 2, fig. 1 ; Plate 3 Cynodesmus nobilis Simpson. Fla. State Geol. Survey, Bull. 10, p. 17, fig. 1, 1932. Two skulls, a nearly complete jaw, and several fragmentary jaws are referred to this species. Except for Amphicyon this is the largest canid yet found in the Florida Miocene. It is very nearly the same size as Cynodesmus thooides Scott. Paradaphaenus tropicalis spec. nov. Type. M.C.Z. 3729 (Plate 1, fig. 2), right half of palate with P4-M2. 6 bulletin: museum of comparative zoology Paratype. M.C.Z. 3714 (Plate 4), left mandible with P4-M2. Horizon and Locality. L. Miocene, L. Arikareean; Thomas Farm, Gilchrist Co., Florida. Diagnosis. One-seventh (15%) smaller than P. nobilis, protocone and metacone of M2 better developed, hypoeone reduced. Two skulls, a crushed rostrum, and some fragmentary jaws are referred to this species. I have placed these two species in this genus on the character of the heel of Mi. Cope (1884 O, p. 900) says of Amphicyon cuspigerus: "The inferior sectorial tooth is characterized by its great robustness; the internal medial tubercle is much elevated, while the principal cusp is short. The heel is wide and basin-shaped, with the inner border as much elevated as the outer." Wortman and Matthew (1899 A, p. 129) made this species the type of a new genus, Paradaphacnus . Scott's figure (1895 C, PI. 1, Fig. 5) shows the entoconid of Mi as very much smaller than the hypoconid. Since the entoconid and hypoconid are nearly equal in nobilis and tropicalis, their affinities are with Para- daphaenus rather than Cynodesmus. The latter genus appears to be restricted to the Upper Miocene and the former is otherwise known only from the Upper John Day. Amphicyon intermedius White Proc. New Eng. Zool. Club, 18, p. 32, pi. 3-4, 1940. Skeletal material and isolated teeth are the only additional material referable to this species. Its affinities are somewhat confused by the combination of advanced and seemingly retarded characters. i Amphicyon longiramus spec. nov. Type. M.C.Z. 3919 (Plate 5), right mandible with P2-M2. Horizon and Locality. L. Miocene, L. Arikareean; Thomas Farm, Gilchrist Co., Florida. Diagnosis. Size and proportions very close to the jaws with skull referred by Matthew (1924 C) to A. sinapius, Pi and M3 single rooted, M3 with groove in the outer side of root but not on the inner side, P4 with heel and accessory cusp, two mental foramina. white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA Measurements Length, condyle to incisors 328 " Pi to M3 153 " Mi 32 " M2 23 diastema C to Pi 18 P,-P2 6 P2-P3 10 Depth of jaw at posterior border of M2 60 This species is about one fourth larger than the preceding, and differs from it also in the double rooted P2 and in the single rooted M3. In dental characters, in size, and in the relative proportions this form is difficult to distinguish from the jaws from the Snake Creek beds referred by Matthew (1924 C) to A. sinapins Matthew, and per- haps does not merit specific designation. One of the peculiar features of this individual is that P4 developed but failed to erupt. The tip is gone from the tooth and perhaps was injured while the tooth was forming. This genus presents many of the characteristics of being a hyper- pituitary Daphacnus. In fact, size is the principal character which separates the two genera. It seems to me that M3 of Daphaenodon is too much reduced for that genus to stand intermediate between Daphaenus and A mphicyon. It is more logical that it should be inter- mediate between Daphacnus and Cy nodes mm. Xothocyon insularis spec. nov. Type. M.C.Z. 3812 (Plate 1, fig. 3), right P4"2. Referred Material. Portion of left maxilla with P4, an isolated un- erupted M1, and a left mandible with Mi. Horizon and Locality. L. Miocene, L. Arikareean; Thomas Farm, Gilchrist Co., Florida. Diagnosis. A large species of Nothocyon, metaconule of M1-2 well developed, hypocone of M1 conical with faint ridge anteriorly and pos- teriorly, hypocone of M2 elongate antero-posteriorly, postero-lateral angle of M2 nearly a right angle and not obtuse as Tomarctns. The lower jaw referred to this species is rather slender and lightly built. The first molar is the only tooth preserved but the distribution of the 8 bulletin: museum of comparative zoology sockets indicate that the anterior premolars were spaced. The cusps of Mi do not differ from those of Tomarctus. Measurements M1 M2 Length x Width 11.3x13.5 7.5x10.8 This species is provisionally referred to Nothocyon because of the spacing of the first and second molars, the quadrangular M2, and the conical hypocone on M1. It may eventually prove to be a primitive species of Tomarctus. It is well advanced toward that genus, however, but in nearly all species of that genus the postero-lateral angle of M2 approaches one-hundred and thirty-five degrees. This species is about the same size as Tomarctus thomsoni, but M1 is much narrower medially. Tomarctus canavus (Simpson) Plate 2, fig. 2; Plate 6 Cynodesmus canavus Simpson. Fla. State Geol. Survey, Bull 10, p. 19, fig. 4, 1932. A crushed skull with P4 to M1 of both sides, a right mandible (M.C.Z. 3628) with P2 to Mi, and a jaw fragment with P4-M1 are referred to this species. This form is only slightly larger than T. thomsoni but does not have the broad medial end of M1 of that species. M2 is missing but the sockets indicate that the postero-lateral angle of this tooth is the greatest in this form of any of the species of Tomarctus. In the lower jaw the premolars are spaced as in Nothocyon and probably were in the upper jaw. Tomarctus thomasi White Plate 7 Proc. New Eng. Zool. Club, 18, p. 94, pi. 14, fig. 3, 1941. A crushed skull (M.C.Z. 3728) with nearly all of the teeth, a palate with most of the teeth, and a right lower jaw (M.C.Z. 3712) with P2 to M2 are referred to this species. This species is not much larger than the preceding (6% as indicated by the molar-premolar length), but the individual teeth are about 12% larger and are crowded together as in T. brevirostris. As near as can be determined from the crushed skull with unworn teeth this species is very close to the skulls figured by Matthew (1924 C) and referred to T. brevirostris. white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 9 It seems reasonable to suppose that Matthew (1930 E) had the species thomsoni and minor in mind when he placed Cynodesmus Cams oorophayus 5 / s PiiocLjon Aelur 'odon Ostec Sorus ^J * i i O Is \ \ Cunoa esmu5 / Ampn Lcuon 0) Paradcji ?/?aenus 1 Tomai -ctus DaphaenoJon -*J s ' Nothc )cuon Daphc 'enus Fig. 1 . Diagrammatic representation of the relationships of a few genera of American Tertiary dogs. Modified from Matthew (1930) and, Vanderhoof and J. T. Gregory (1940). between Nothocyon and Tomarctus in his phylogenetic arrangement. Since these and other species have been removed from Cynodesmus, 10 bulletin: museum of comparative zoology it has become a monotypic genus. Certainly the affinities of Cynodes- mus thooides Seott are closer to Daphacnodon than to Nothocyon. In the light of the recent taxonomic changes (Vanderhoof and J. T. Gregory, Univ. Calif. Bull. Dept. Geol. Sei., 25, p. 160, 1940; and White, Proc. New Eng. Zool. Club, 18, p. 95, 1941.) it seems desirable to make some modifications in the graphic representation of the relation- ship of these genera. See Text fig. 1, which is modified from Mat- thew (1930 E) and Vanderhoof and J. T. Gregory (ibid, p. 145). Two genera, Paradaphacniis and Nothocyon, of the fossil dogs found in this deposit are restricted to the Upper John Day and the Lower Rosebud of the great Plains. Two species of Tomarctus are present, one T. canavus, exhibits a less advanced character in the small size and in the spacing of the premolars, and the other appears to be more advanced and very similar to T. brevirostris. One of the species of Amphicyon (A. longiramus) is very advanced and very close to A. sinapius of the Middle and Upper Miocene. The affinities of the other species are not altogether clear. The specimen referred to Daphaenus appears to be somewhat degenerate and therefore worthless for corre- lation. On the whole the Canidae exhibit a basal Miocene aspect. ARTIODACTYLA While there is considerable variety among the artiodactyls (eight species and as many genera) found in these deposit, certain groups are conspicuously absent (Entelodontidae, Merycoidodontidae, and Agriochoeridae). Since they are so abundantly represented in the Miocene deposits of the Plains it seems reasonable to suppose that they did not inhabit Florida at this time. Of the artiodactyls present only the Cervidae and the Protoceratidae are represented by more than one specimen in each species. TAYASSUIDAE Floridachoerus olseni White Proc. New Eng. Zool. Club, 18, p. 96, pi. 14, fig. 4, 1941. A few isolated teeth are the only additional material of this species found. It appears to be more advanced than Desmathyus and less so than Prosthcnops. WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 11 CAMELIDAE OXYDACTYLUS FLORIDANUS Simpson Fla. State Geol. Survey, Bull. 10, p. 35, figs. 20-21, 1932. No additional material certainly referable to this form has been found. Concerning the affinities of this species Simpson (1932 D, p. 16) says: "Oxydactylus fioridanus belongs to a lower Miocene group, so far as its affinities can now be read, and appears to be a rather advanced member of that group. The Midway camels seem to be somewhat more progressive, although the evidence is very poor, and this may be illusory". Paratylopus graxdis White Proc. New Eng. Zool. Club, 18. p. 33, pi. 5, 1940. No additional material certainly referable to this species has been found. Isolated molars and premolars are rather common, but it is difficult to be certain whether they belong to this form or to Oxydac- tylus. This is the largest species of this genus known, but its affinities are not altogether clear. HYPERTRAGULIDAE There are three genera of this family represented, all of which seem to be peculiar to this deposit. The smallest and more nearly normal is represented by a fragment of a right mandible with Dp2-3 and Mi. It seems better to withhold a specific diagnosis until better material is obtained. Hypermekops genus nov. Genotype, olseni spec. nov. Diagnosis. A large brachyodont hypertragulid with three incisors in the premaxillary, fourth premolar and molars similar in form to those of Leptomeryx, P3 three rooted and probably with a median spur, P2 double rooted, elongate entero-posteriorly and without median spur, I1 to P1 caniniform and slightly recurved, I1 largest. Hypermekops olseni spec. nov. Type. M.C.Z. 3711 (Plate 8), a skull containing I1"2, P2 and 4, and M1"3 of the right side, and l1 and P4 to M3 of the left side. 12 bulletin: museum of comparative zoology Horizon and Locality. L. Miocene, L. Arikareean; Thomas Farm, Gilchrist Co., Florida. Diagnosis. Same as generic. The skull was crushed flat when found but has been expertly re- stored for exhibition by Mr. Russell Olsen. Fully realizing the im- portance of this specimen, he performed this feat without disturbing the palate from the condition in which it was received at the museum. The bone of the anterior end of the snout was sufficiently dense and heavy so that the crushing did not shatter the bone but mashed it down in clean breaks which fitted together perfectly when the matrix was removed. Consequently there can be no doubt about the restora- tion of the end of the snout. There were good contacts for the pieces of bone of the dorsal side of the face all of the way back to the frontal crest. Both postorbital processes are complete so there can be no doubt about the position of the orbits. The occiput is moderately well preserved, and also the inferior and the antero-inferior borders of the orbits. Taking these into consideration there can be little doubt about the height of the frontal region of the skull. The nasals are separated from each other and from the adjacent bones by suture. The exact nature of the fronto-nasal suture cannot be determined, but it appears to be W-shaped with the apices directed posteriorly. The frontals are separated from each other by suture. The two halves of the palate are separated by suture, but no trace of the maxillo-premaxillary suture can be found. The first incisor shows wear on the antero-medial and posterior sides. This indicates that it occluded with the first and second incisors of the lower jaw, which would enable the animal to get a firm grip on some fleshy part of a plant, such as roots and tubers, and rip it loose. The closed maxillo-premaxillary suture lends support to this hypothesis. Also the bones of the snout have the same dense, polished appearance that those in the snouts of pigs and peccaries have. However, the tip of the snout of this form seems to be too lightly constructed for any strenuous rooting. The second incisor is smaller than the first and strongly recurved. It shows no wear except on the tip. The third incisor, the canine, and the first premolar are missing, but, judging by the sockets, they were about half the size of the first incisor. Mr. Olsen took very careful measurements of the palatal side of the skull before beginning the restoration. The distances from the tip of the snout were taken at the tip of the tooth. WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 13 Condylo-basal length Condyles to M3 Anterior border of orbit to tip of snout M3 to tip of snout M1 " it a a P2 " t a tt Pi « t (( (I C " c a it I3 " t a a I2 " t tt tt I1 " i a it To these may be added: Length P2 to M3 381 130 275 251 214 185 134 114 99 77 31 81 Floridatragulus dolichantherius White Proc. New Eng. Zool. Club, 18, p. 34, pi. 7, 1940. A few isolated teeth are the only additional material referable to this species which has been found. Since both Hypermekops and Floridatragulus have extenuated snouts, it seems reasonable to suppose the former to be the ancestor of the latter, which has the longer snout. However, until we find corre- sponding parts of both forms we cannot be sure. The presence of all three upper incisors in Hypermekops indicates that it developed from some Upper Eocene hypertragulid which was able to find a satisfactory ecological niche here. PROTOCERATIDAE Syndyoceras australis White Proc. New Eng. Zool. Club, 18, p. 97, pi. 15, 1941. Although two additional mandibles of this form have been found they do not greatly increase our knowledge of it. As with Protoceras this form seems to be quite variable, and in view of the kinship of the two genera the differences presented are probably sexual rather than specific. If the premolar-molar index is any indication of the degree of ad- vancement of a species, australis is slightly less advanced than cooki. 14 bulletin: museum of comparative zoology The indices of the two species are: australis, 28/58 — 49%; and cooki, 32/61 — 52%. As far as we know this genus is restricted to the Lower Miocene. CERVIDAE Machaeromeryx gilchristensis White Proc. New Eng. Zool. Club, 18, p. 97, pi. 14, fig. 5, 1941. The first and second upper molars with a fragment of the maxillary, and a lower jaw with well worn teeth are the only additional speci- mens of this species which have been found. The longer premolar series would seem to indicate that this species was more advanced than M. tragulus. This genus appears to be restricted to the Lower Miocene. Parablastomeryx floridanus White Proc. New Eng. Zool. Club, 18, p. 34, pi. 6, 1940. Three additional lower jaws of this species have been found. Its size is about that of P. gregorii. For comparison I have copied the premolar-molar indices given by Frick (Bull. A.M.N.H., 69, p. 227, 1937) for the species of this genus. »5 1 ** ^ ♦ 2* > 5 + 5 1 "3 9JDJDa39 3UOOO)Ojd 1 0. § c: I 5 &, 0> 8 X X o o "3 .C; — > X X X X X X X X <§- 8 J 1 .5 X X 5 X X X 5; 0 5 "5 5 *** "> ^ •u u e 1 1 ^0 X 1 X X X X X <§- X X c: 0 0 X X X * Q. 5 1 .c: ■■s X X ^ M ^ X X X X X s t X X •*> 4. X u2 X X -c: ^ OQ *> 5- K ^o d- d~ 25 CO i- ■-0 * d- ^ iQ ^Q *- 5 ^ ^ *■ CO CO ^ ^ CQ •^1 "> ^ co O ^ (\1 ^ O °0 0 °0 00 O ^0 * t ^ "0 O °0 ^> °0 °0 °0 °0 °o "3 8 IS vo <5- °0 1 CM rO CO ^0 CM O 00 <0 CO *o vo Is. CO 0- s. CO 5 03 CO °o O *■ CO CO 55 CO 0 CO °0 Fig. 2. Table of (Hay). measurements and variations of Parahippus blackbergi 18 bulletin: museum of comparative zoology In the lower jaw the posterior lobe of the third molar is very little relatively than in the next species of Parahippus found in this deposit. Four of the lower dentitions have cement on the molars. Protocone. In most specimens there is only a constriction between the protocone and the protoconule, but in Nos. 3820, 3831, and 3815 the tip of the protocone is separate in a few of the teeth, usually the premolars. This is also the case in some of the isolated teeth. Hesse (manuscript) reports that he finds the same condition in the specimens from Garvin Gully. Milk Teeth. One specimen (M.C.Z. 3840, Plate 10, figs. 2 and 3) still has the milk teeth. The protocone is large and deeply constricted from the protoconule. The metaloph is connected to the ectoloph but shows no plications. The hypoloph is well developed and projects into the postfossette. The hypostyle and posterior cingulum are well developed. Plihypostyle. Ever since the name plihypostyle was first applied to a feature in the postfossette of horse teeth, there has been considerable confusion regarding the homologies of structures bearing that name. It appears to have been used, at one time or another, for any structure in this area which could not be identified as the hypostyle. Obviously the plihypostyle of Parahippus cf. ncbrascensis (Stirton, Journ. Mamm., 22, p. 434, fig. 3, 1941) is not the homologue of the element bearing Protoconule Protocone MQZ3922 Hupo/oph Hypodtu/e Hi T ocone Fig. 3. Third upper premolar of Parahippus blackbergi (Hay) showing hypoloph. x 2. that name in Neohipparion cf. eurystyle (ibid, p. 435, fig. 7). By extreme good fortune the fossil Equidae from the Thomas Farm demonstrate the development of the principal features in this area. Some of the specimens of P. blackbergi (Nos. 3820, 3815, and 3829), and a number of isolated teeth, show, on unworn or little worn premolars, a high thin white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 19 ridge (M.C.Z. 3922, Text fig. 3) extending from the hypocone to the posterior cingulum near the metastyle. By all of the rules of tooth terminology this ridge is the hypoloph and makes up the posterior half of the metaselene. In two of the specimens (Nos. 3815 and 3829) the metaselene is complete on one or more of the molars. In its simplest form (P2 of No. 3815) the hypostyle is a small conical tubercle on the posterior cingulum. However, in most cases it takes the form of a spur, projecting postero-medially from the inner end of the hypoloph. This condition is especially well demonstrated in the milk teeth of Mcrychippvs paniensis and to a lesser degree in the milk teeth of M. primus. The hypoloph, in the form it takes in the premolars of P. blackbergi, is present in the molars of 'P. Iconcnsis and the less progressive specimens of M. gimteri, but this is not true of the premolars of P. Iconcnsis. In the more progressive specimens of the latter species the postfossette wall of the hypoloph often bears a plication near the postero-lateral end. This appears to be homologous with the plihypostyle of Neohip- parion cf. Eury style (ibid, p. 435, fig. 7) and for the sake of clearness and compatibility, I propose that this name be restricted, so that it will apply only to plications on the inner wall of the hypoloph. While this may not be in strict accordance with the customary practice in mor- phological names, it will make the terminology of the upper teeth more consistent with that of the lower. This species presents many variations which are departures, in the direction of Parahippus, from a simple pattern similar to Miohippus. It is possible to find all of the generic characters of the teeth of Para- hippus in this series of specimens. The greatest number of characters found in combination in one individual (Xo. 3S31) is four, and many have three. These characters are: * 1. Well developed crochet. 2. Plications on the anterior and posterior walls of the metaloph. 3. Protocone separate at tip. 4. Hypoloph and posterior cingulum closing the postfossette. 5. Cement on molars. 6. Height of unworn M2 at paracone only slightly less than external length. These characters, occurring in combinations of as many as four in one individual, indicates that this species stands genetically very close (and is probably ancestral) to some of the earlier species of this genus. The simple and stable pattern of the cheek teeth presented by the Upper Miocene species of Archaeohippus and the more complex and 20 bulletin: museum of comparative zoology highly variable character of the teeth of this species presents us with two possibilities. Either the Florida stock retrogressed, or, the Plains were unsuited ecologically to this stock and they remained in the vicin- ity of the Gulf. Owing to the nature of the deposits in that area they are recorded only in Garvin Gully. In view of the Miocene geography of the Florida region (see that section), this stock can, at best, be only remotely related to that of the Plains, probably only in that they have Miohippus for a common ancestor. Consequently, to leave the Florida and Plains stocks in the same genus is a purely artificial classification, which defeats one of the primary purposes of taxonomy, in that it should be an expression of the genetics of groups of animals or plants. According to the available evidence this form is a very primitive species of Parahippus, bridging the gap between fhat genus and Miohippus. The weight of the characters lean more toward Parahippus than Miohippus, and it would not simplify matters any to name a new genus for this intermediate and highly variable form. Parahippus barbouri spec. nov. Type. M.C.Z. 3646, (Plate 11), a crushed skull which has been re- stored for exhibition, M3 unworn. Paratype. M.C.Z. 3814 (Plate 12) right lower jaw with P2 to M3. Referred Material. M.C.Z. 3736, upper dentition lacking left M3; and No. 3742, a well worn upper left dentition. Horizon and Locality. L. Miocene, L. Arikareean; Thomas Farm, Gilchrist Co., Florida. Diagnosis. A small Parahippus with a moderately heavy coat of cement on all cheek teeth, plications on anterior and posterior walls of metaloph, crochet simple and usually not in contact with proto- conule, protocone joining protoconule on all teeth by the time M3 has begun to receive wear, postprotoconal valley open except on M1, post- fossette may or may not be open on M1-2, external styles well developed teeth subhy sidont. Measurements Number 3646 3736 3742 Length, P2~M3 87 88 82 ' ' p2— 4 45 46 42.5 " M1"3 42 42 39.5 Width P4 17 17.5 17 " M1 17 18 16.5 WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA • 21 The limited amount of material referable to this species, each show- ing a different stage of wear, does not give us any data on the amount of variation in the enamel pattern. The enamel pattern of this species is very close to that of P. crenidens (Scott) from the Deep River, U. Miocene of Montana and to P. coloradcnsis Gidley from the Pawnee Creek Beds of Northeast Colorado. So similar are their enamel pat- terns that it seems logical to suppose that the Florida stock persisted till near the close of the Miocene with only a slight increase in size. Parahippus leonensis Sellards Plate 13, figs. 1 and 2 Eighth Ann. Report of Fla. State Geol. Survey, p. 83, pi. 11, fig. 7, pi. 13, fig. 2-3, 1916. Merychippus vellicans Hay, Proc. Biol. Soc. Wash., 37, p. 7, pi. 1, figs. 18-19, 1924. At first I referred this material to Parahippus vellicans (Hay), be- cause most of the specimens were nearly identical with Hay's types. However the last three skulls which have been cleaned have convinced me that Sellard's type is an unusual varient of this population and consequently the species must be known as P. leonensis. Only one or two of the teeth bear the peculiar type of crochet found in the type of P. leonensis, nor is it always on M1, but may be found on any tooth except P2. Mr. C. J. Hesse wrote, after examining a small series of teeth sent him; that in his opinion the Thomas Farm and Garvin Gully popula- tions were conspecific. Thirty-three specimens, represented by the upper cheek tooth series of at least one side, are referred to this species. A nearly equal number of lower jaws probably belong here. However, none of the lower jaws were found associated with the upper. This species appears to embrace as wide a range of variation as P. blackbergi. Several of the specimens exhibit characters of size and enamel pattern suggestive of P. barbouri. It is hoped that further excavation will complete this gap in the record. The other extreme is with difficulty distinguishable from M. gunteri. At first, I divided this series of specimens into three groups on the basis of the extremes of enamel patterns. For a period of about three weeks I reviewed them once or twice a day. Each time I shifted a few specimens from one group to another. Since it was impossible to make any division that would last over night, I concluded that only 22 bulletin: museum of comparative zoology one species was represented. The individuals of this species seem to represent three groups in the population; a conservative, a progressive, and an aberrant. The intergradation between these three groups seems to be complete. Cement. The amount of cement varies from a very heavy coat (more than M . •primus) to a very thin film. The usual amount seems to be a little less than M. primus. The third molar always has the most cement and the first has the least. Crochet. The crochet is usually T-shaped with the crossbar of the T abutting against the protoconule. Sometimes it is only a V-shaped projection on the metaloph with the apex of the V directed toward the protoconule. Or, it may be a long narrow loop projecting between the protocone and the hypocone. Metaloph. Unworn or slightly worn teeth bear plications on the anterior and posterior walls of the metaloph. Those on the posterior wall are smaller and fewer than those on the anterior, and soon disap- pear with wear. The plication nearest the metaconule is the strongest and may persist as long as those on the anterior wall. The plications on the anterior wall are usually three or four in number, with those nearest the crochet the strongest. M3 and P2 usu- ally have only one or two plications on the anterior wall. Hypoloph. On the molars and sometimes on P4 the hypoloph ex- tends from the metaconule to the posterior cingulum near the meta- style. This is the same condition found in the premolars of P. black- bergi. In the premolars the hypoloph projects into the posterior part of the postfossette and does not reach the posterior cingulum. -Its direction is more nearly lateral than postero-lateral as in the molars. A plihypostyle may or may not be present. It is more common on the molars than on the premolars. Hypostyle. The unworn hypostyle is usually covered with cement, so that its shape cannot be determined. After a small amount of wear it is usually triangular in the molars and elliptical in the premolars, with the long axis parallel to the hypocone. It often appears as a spur directed postero-medially from the inner end of the hypoloph. In the more advanced specimens the hypostyle has gained the ascendency in growth over the hypoloph and receives wear first. It has the form of a long narrow loph paralleling the hypocone. The inner end usually projects into the postfossette. This projection cannot be considered a plihypostyle because of its origin as demonstrated by this material. Both features occur in some specimens. This condition is prophetic of that found in M. gunteri on all teeth except M3. WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 23 Protocone. The protocone is usually joined to the protoconule by the time M3 has begun to receive wear. It usually, but not always, displays a spur. There seems to be a positive correlation between the amount of cement and the length of time which the protocone is separate. In the two specimens with a very heavy coat of cement the protocones are separate even though M3 has received considerable wear. Milk Teeth. A number of specimens show the milk dentition in various stages of wear. No. 3759, with P1 and M1 just beginning to receive wear, displays the characters of these teeth the best. The enamel pattern of the milk teeth is very similar to that of the permanent premolars. The crochet is well developed but usually does not abutt against the protoconule, more often it is broader at the base than at the outer end. The metaloph is connected to the ectoloph and has plications on both walls. Those on the posterior wall dis- appear first. Hypoloph and hypostyle are like those of the permanent premolars. Cement absent except for a thin film on the exterior base of two specimens. The milk teeth of this species are very similar to a cast of the type of Parahippus eognatus Leidy and the permanent teeth are very close to the type of P. brcvidens (Marsh). Certainly these three species are very close genetically and possibly should bear the same name, in which case, they should be known as P. eognatus Leidy, since that is the earliest available name. The scarcity of the remains of eognatus and bremdens would lead one to believe that this stock was well past its prime in the Upper Miocene. It does not seem unreasonable that a stock which received its inception in the Lower Miocene should persist with diminishing abundance till near the close of the Miocene. Merychippus gunteri Simpson Plate 13, fig. 3 Bull. Amer. Mus. Nat. Hist., 59, p. 165, fig. 10, 1930. This species unquestionably arose #from Parahippus leonensis Sellards. The transition between the two species is well represented by complete dentitions. Fortunately it is still possible to set up an arbitrary rule which will distinguish the conservative members of this species from the progressive members of the preceding one. This rule is:— The crochet must have joined the protoconule on M1-2 by the time wear has exposed the principal cusps on M3 before the speci- 24 bulletin: museum of comparative zoology men can be placed in M. gunteri. At this time both species have a height of tooth at the mesostyle on M2 of 10 mm. When the teeth of the progressive specimens of P. leonensis are worn so that M2 has a height of 7.5 mm at the mesostyle they display most of the characters of M. gunteri. None of the unworn teeth of either species have a height of crown to exceed 15 mm at the paracone. Unfortunately the material representing this species is not plentiful ; a badly broken skull with good teeth, a complete upper dentition of both sides, three specimens with five cheek teeth in series, and three with three or four teeth. Cement. x\ll specimens have a uniformly heavy coat of cement on the outside of the tooth, about the same as the most advanced speci- mens of P. leonensis. M1 always has the least amount of cement in the fossettes and M2 is next. In M3 and the premolars the fossettes are usually nearly filled. The postprotoconal valley has less cement than the fossettes. Crochet. On the molars the crochet has joined the protoconule before the tooth has received much wear, often on M1-2 before M3 has erupted. A greater amount of wear is necessary in the premolars for these two structures to join, and apparently they never join in P2. Usually there is only a single plicabellin and pliprotoconule but one specimen displays two of each on M1. Hypoloph. The hypoloph and hypostyle are very little different from those in the premolars of P. leonensis except on M3 in which, on most specimens, it exhibits its earlier characteristics. A plihypo- style is often present on the premolars but is not so common on the molars. Metaloph. The plications on the metaloph do not differ materially in size and distribution from those on P. leonensis, except that they seem to persist longer in those species. One rather unusual specimen shows five plications on the anterior wall and four on the posterior wall of M1. Protoeonc. The protocone always has a spur except in the early stages of wear on P2. It usually joins the protoconule on all teeth by the time M3 has begun to receive wear. Milk Teeth. An isolated Dp2 seems referable to this species. The tooth has received considerable wear but not enough to obliterate the details. No trace of cement can be found. The crochet has joined the protoconule and there is a suggestion of a plicabellin and a pliproto- conule. There is one large and one small plication on the anterior wall of the metaloph. The posterior wall is smooth. The hypoloph is white: lower miocene mammal fauna of Florida 25 well developed and separated from the hypocone by a thin line of enamel. The hypostyle is small and triangular in outline with the apex directed toward the inner end of the hypoloph. The protocone bears a spur and is separated from the protoconule by a thin line of enamel. It would be comparatively easy to derive the enamel pattern of most of the species of Merychippus from that of this stock. The sim- plest pattern exhibited is only slightly more complex than that of M. primus and the most complex is about equal to that of M. sphenodus (Cope) and M. calamarius (Cope). The pattern of most of the speci- mens is about the same as that of M. secundus Osborn (if that is a valid species). However, if, as Stirton (1940, Univ. Calif. Bull. Dept. Geol. Sci., 25, p. 181, Footnote 12) suggests, M. secundus, tertius, and quint us are synonyms of M. primus, the size and height of crown become the principal differences between gunteri and primus. This difference is sufficiently great to justify retaining them as distinct species. Simpson (1932 D, p. 27, Footnote 5) suggests that M . gunteri and primus arose from different species of Parahippus. In the light of this material I believe it more likely that the latter is the descendant of the former. Merychippus westoni Simpson Bull. Amer. Mus. Nat. Hist., 69, p. 164, fig. 9, 1930. In the material from the Thomas Farm I have not been able to identify with certainty this species. In the isolation of the protocone it seems to be more advanced than gunteri and probably represents a later stage in evolution. These horses are less advanced than those of the Middle and Upper Miocene, although their closest relatives are found in deposits of those ages. They appear to have very little in common with the Lower Miocene horses of the Great Plains. On the whole they are of very little help in correlating the age of this deposit. Development of Tooth Form in Horses as Indicated by this Material x\lthougb this series is far from complete, there are a sufficient number of stages so that a moderately clear picture can be obtained. Separation of Protocone. In P. blackbcrgi the protocone appears to be separated from the protoconule first on P3~4. The protocone is usually deeply constricted from the protoconule on P2 and M1 but the tips are seldom separate. It may be separate on M1 and not on P2. 26 bulletin: museum of comparative zoology Unfortunately there are no specimens showing the protoeone becoming separated on M2-3. In the other species of Parahippus and in Mery- chippvs the tip of the protoeone is separate in the early stages of wear. Cement. The cement first appears on M1 or 2 in P. blackbergi. In the cases in which it is present only on M2-3 1 cannot be certain that it never had been on M1. It is thought to be present on the premolars of No. 3820 but it is difficult to be certain. In the other species of Para- hippus and in Meryckippus M1 always has the least cement and M3 has the most. Among the premolars P2 has the least and P4 the most but none as much as M3. There seems to be a positive correlation be- tween the amount of cement and the order of tooth succession. Hypoloph and Hypostyle. The development of these two features were fully treated in the consideration of P. blackbergi. Crochet and Plications on the Metaloph. The form of these features are fairly well catalogued under the various species. There is, how- ever, a plausible explanation of their mode of formation based on the embryological principal of Unequal Growth. Both embryology and the observations of unerupted teeth show that the durable layers of the teeth are deposited from the crown to the base. The general form of the tooth is determined by the meso- dermal papillae which outline in a general way the principal cusps and lophs, which grow down (or up in the lower jaw) to meet the dental germ and is enveloped by it. As the dental germ (enamel depositing cells) spreads over the sides of the lophs into the fossettes the cells must divide rather rapidly to supply the necessary amount of en- velope. If the cell division is more rapid than necessary, or if it con- tinues after the loph is covered, it creates more surface than there is space to accommodate it. This internal pressure, pushing against the already established cusps, which act as buttresses, causes the surface to buckle and be thrown into folds. The crochet, after the metaloph has joined the ectoloph, occurs in the logical place for the surface to buckle from the internal pressure set up by the rapidly multiplying cells of the dental germ layer; i.e.; at the apex of the curve of the metaloph. The location of the "anticrochet" is determined in the same way. That the secondary folds are formed subsequent to the formation of the principal features is indicated by the fac1> that they always occur in the areas of thin enamel. There are a number of un- erupted teeth in the material from Florida in which the deposition of enamel is complete and which contain no dentine on the inside of the tooth. This indicates that the details of tooth form are determined by the proliferation of the enamel depositing cells. white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 27 Union of Crochet and Protoeonvlc. The embryology of the tooth permits us to postulate the steps through which the crochet becomes united to the protoconule. In some specimens of P. blackbergi the crochet abuts against the protoconule. This is true of nearly all of the specimens of the other species of Parahippus found in Florida. The protoconule must have been well established by the time the crochet began to form. This resulted in the latter flattening its end against the protoconule. To create additional space for the over abundant enamel depositing cells the crochet became constricted at the base by the process mentioned above. By this time the crochet is T-shaped. The crossbar of the T in the crochet of Parahippiis becomes the plicabellin and the pliproto- conule of Mcrychippus. Eventually the crochet came abreast of the protoconule in its order of development and these two areas of growing tissue met, fused, and split in the same manner as the foetal mem- branes. I have been unable to find a statement of this principle in any of the textbooks of embryology at hand but I think the following will express the idea sufficiently well; "When two outpouchings of growing tissue from the same germ layer meet, first there is a fusion of the two surfaces, then a fission in a plane parallel to the axis of juncture but at right angles to the plane of fusion. It seems highly probable that this same process was involved in the union of the metaloph and ectoloph, and also in the isolation of the protocone in Hipparion (sensu latu). SUMMARY Eighteen genera and twenty-two species of mammals have been identified from the Thomas Farm in Gilchrist County, North Florida. In order to determine the age of this deposit the closest relatives of these species were sought in the deposits of the Great Plains. The Artiodactyla present a decided Lower Miocene aspect. Two of the genera are restricted to deposits of that age on the Plains. Two genera of one family (Hypertragulidae) seem to be peculiar to the deposit in North Florida. Among the Carnivora are two genera which did not persist on the Plains beyond the Upper John Day - Lower Rosebud age. There is one species in each of two other genera which are very closely related to plains species in the same genera of the Late Middle and Upper Miocene. The Equidae are all progressive and their only relatives are found in the Late Middle and Upper Miocene deposits of the Plains. Since the vertebrates give such a paradoxical correla- tion, other sources were examined for data to aid in determining the 28 bulletin: museum of comparative zoology age of this deposit. Overlying the deposit in which the mammalian fossils occur is an erosional remnant of what is believed to be a mem- ber of the marine Hawthorne formation. However, no identifiable in- vertebrates have been found in it. Invertebrate fossils, as well as cobblerock, from the Ocala and Suwannee limestones are found mixed with the vertebrate fossils. It would seem then that the mammalian remains were buried during the interval between the deposition of the Suwannee limestone and the deposition of the upper part of the Haw- thorne formation. An examination of the geologic map of Florida shows that this local- ity was near the northern end of an island during the Lower Miocene. This gives a plausible explanation of the anomalies in the correlation of the vertebrate fauna. The forms in this fauna closely related to the Upper Miocene forms of the Plains developed on the island and were unable to escape to the mainland till after the Early Middle Miocene. WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 29 2. GEOLOGY LOCAL DETAILS Plate 14 The Raeford Thomas Farm is located 8 miles north of Bell in northern Gilchrist Co., Florida. It is near the eastern edge of the watershed which separates a series of small lakes on the east (which drain by an unnamed creek into the Santa Fe River) from the Su- wannee River on the west. On this watershed frozen (or fossil) sand dunes are the principal topographic features, with sinkholes running a close second. The top soil, in some places to a depth of many feet, is yellowish sand which may be the result of the weathering of the Hawthorne formation. The pit from which the fossils are being taken is on the eastern edge of a circular depression about 250 feet in diameter and about 20 feet deep. The form of the depression with its gently sloping sides and shallow depth is closer to that of a "blow out" between sand dunes than it is to a sink hole. A test hole, dug with a six inch auger, en- countered a rocky stratum at eight feet. The data obtained was not sufficient to determine whether this stratum was the Ocala limestone or a boulder bar similar to the one encountered in the pit. Although the evidence concerning the origin of this depression is not conclusive, there is no reason to suppose that it played any part in the deposition of the sediments which carry the mammalian remains. For additional information concerning the area around the pit I am greatly indebted to the Florida State Geological Survey for contribut- ing the time of Mr. Clarence Simpson, who spent two days boring testholes around the excavation. He took numerous samples from the test holes which were analyzed in the laboratory of the Survey. The main part of the pit is about 100 feet long and about 60 feet wide (Plate 14, fig. 1). The pit is about 14 feet deep in the deepest part. To facilitate the removal of the refuse earth a narrow trench, wide enough to permit a mule with a scrape to pass, was dug from the pit to the deepest part of the depression mentioned above. The surface of the area from which the fossils are taken is covered by loose, yellowish sand which varies in depth from a few inches to a few feet. The color and texture of this sand suggests that it may be the result of the disintegration of the calcareous sandstone members of the Hawthorne formation. Below the loose sand is a layer of joint 30 bulletin: museum of comparative zoology clay which varies in thickness from one to about four feet. In the area west of the boulder bar (Plate 14, fig. 4) the clay grades into the lime sand below it. It was here that the Florida State Geological Survey dug its pit and the Museum of Comparative Zoology started its excavation. A few species were found in this part of the excavation which have not been found in the deeper part. These forms are; Anchitheriwrn clarencei, Merychippus gunteri, and Mephititaxus ancipidens. The only identifiable specimens of Oxydactylus floridaivus, Paratylopus grandis, and Floridatragulus dolichanthcrvus were found here. However loose teeth and other fragments were found deeper in the pit. The joint clay contains much broken bone and loose teeth. A few horse skulls were found in this layer but were very badly broken although not badly crushed. At the eastern end of the pit, between the surface sand and the joint clay, is about three feet of cream colored, pumice-like sandstone with brown spots. Lithologically it resembles the upper members of the Hawthorne formation exposed in the Devil's Mill Hopper near Gainesville and that in a road cut between the Seaboard Railway Sta- tion at Gainesville and Bivin's Arm. Sandstone similar to that found at the pit, is exposed in road cuts along the highway north of Bell. The test holes, bored by Mr. Clarence Simpson, indicate that this sandstone once covered the area now being excavated. If the correla- tion of this sandstone is correct, this cycle of fluvial deposition was brought to a close by the invasion of the Hawthorne sea. Below the Joint clay is a layer of clay balls, which as near as could be determined, are eight to twelve inches in diameter. In some places they are piled four or five deep. In reality they are a mass of bone fragments and teeth held together by clay. Occasionally a nearly complete leg bone or jaw is found, but is usually very badly broken. There is a strong possibility that the bone fragments and isolated teeth are reworked material. At the west end of the pit, lying partly below the clay balls and partly below the joint clay, is a lens of lime sand whose greatest thickness is about three feet. The lime sand grades in the joint clay where the two are not separated by the clayballs. Also there is a lens of lime sand below the thin southwestern edge of the boulder bar (Plate 14, fig. 3). Most of the lime sand is moderately coarse but some of it is fine enough to be classed as silt. Most of the bones found in the lime sand are whole, but often they are so soft that it is im- possible to save them. Most of the specimens obtained by the Florida State Geological Survey came from the lime sand above the boulder WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 31 bar. Merychippus gunteri and advanced specimens of Parahippus leonensis occur here. Lying immediately below the beds mentioned above and extending diagonally across the excavation (Plate 14, fig. 1) is a bed of boulders which vary in size from two or three inches in diameter to as much as fifteen in the largest ones. However, the majority have a diameter of six or seven inches. The interstices between the boulders are filled with gravel and lime sand. All of this material appears to have been derived from the Suwannee and Ocala limestones. The material is entirely un- sorted. The bed is thickest at the middle and thins out at the edges. This bed contains many horse skulls which are hopelessly crushed. Some are moulded around boulders so that it is impossible to save them. Many invertebrates characteristic of the Suwannee and Ocala limestones are found in the boulder bar. Below the boulder bar the excavation has penetrated about six or seven feet into a bed of laminated bluish clay. Many of the laminae are separated by a thin layer of clean, white quartz sand, while others are separated by a thin layer of silt. There are local pebble layers which may be two or three inches thick. The bedding is very lenticular with most of the lenses only a few feet across. The preservation of the bone is the best in this part of the excavation. Although the skulls are crushed the bone is quite firm so that restoration is possible. The lack of sorting of the material which makes up the boulder bar, and the lenticular nature of the laminated clays below it, indicate that these sediments are not of marine or lacustrine, but of fluvial origin. The general outlines of the history of this stream are fairly obvious. The initial stage was a period of erosion during which the stream scoured out its channel in the soft Ocala and Suwannee limestones. Next the force of the stream was lessened so that it began to silt up its channel, burying the remains of the animals which perished along its course. Later the stream was rejuvenated and its carrying capacity increased to the extent that it was able to transport the large boulders which make up the boulder bar. Still later the transporting power was again reduced so that it was able to carry only fine sand and silt. Eventually the lower reaches of the stream were drowned by the ad- vance of the Hawthorne sea. The data for limiting the period of time represented by this deposit are the presence of Ocala and Suwannee invertebrates in the deposit and the Hawthorne formation lying above it. Thus this period of time began after the deposition of the Suwannee limestone (Upper- 32 bulletin: museum of comparative zoology most Oligocene) and came to a close before the end of Hawthorne time (Late Lower Miocene). Unfortunately there are no data concerning the duration of the period of channel cutting. The Ocala and Suwannee limestones are poorly consolidated and would erode very easily. If the structure of the Central Florida Dome at this time was at all similar to what it is today the gradient would be about four feet per mile. This would be sufficient gradient to cut a considerable channel in a relatively short time. The events which caused this stream to silt up its channel are not at all evident. Although Mansfield (1937, p. 42) presents evidence that the Tampa was a transgressing sea and that the upper zones overlap the lower, I believe this part of the stream was too far from the shoreline to have been affected. In view of the purity of the Ocala and Suwannee limestones this stream must have drained a very large area to have accumulated the amount of clay found in this deposit. Mansfield (1937, p. 44) insists that there was a period of uplift at the close of Tampa time. As near as can be determined from the distribution of the Tampa and Hawthorne sediments, the western edge of the Florida Plateau was elevated at this time. It seems logical that this would increase the size of the drainage basin and lengthen the stream. The increased drainage basin would increase the head of water and the transporting power so that it would be able to pile up the boulders found in the boulder bed. As the Hawthorne sea ad- vanced the stream again lost its power to transport boulders and cobble rock, and deposited sand and clay in its channel. Eventually this portion of the stream was drowned by the Hawthorne sea. PALEOGEOGRAPHY An examination of the geologic map of Florida shows that the area in which the mammalian fossils are found was undoubtedly an island during Lower Miocene time. In order to get a better concept of the sequence of events it seemed desirable to compile maps of the Early Tertiary formations of the southeastern United States. In these maps the solid lines represent the known limits of the outcrops and are not intended to represent old shore-lines. The dashed lines indicate the probable limits of the formation in areas in which it is buried under younger formations. Although I have covered most of Florida by auto during the past three winters, I have taken the data for these WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 33 maps from the more recent literature. For convenience I have in- cluded a correlation chart of the Oligocene and Miocene taken from Cooke (1935), except the Alabama Miocene which is from Semmes (1929) after Cooke. S. CAROLINA GEORGIA FLORIDA ALABAMA Eastern Western Eastern Western K Z in o a o t-< O 5= 3 - < Shoal River Oak Grove Sand Hattiesburg Clay o o Hawthorne Hawthorne Hawthorne Tampa Is. Chipola Marl Catahoula sandstone Tampa Is. Tampa Is. Flint River Flint River Suwannee limestone Flint River Flint River Chichasawhav Marl Absent? Repre- sented Absent? Bucatunna Clay z u I— 1 0 Absent Absent Glendon limestone Absent Glendon limestone Absent? Absent? Absent Marianna Is. Marianna limestone /Red /Bluff / lay z Cooper Marl Barn- well Barn- well Ocala Is Tivola tongue of Ocala Is. Ocala LS. Ocala LS. c o m 0) .5 - o O d c3 o M-l d o u 03 1-5 Yazoo Clay K • r f f c t J Santee Is. C 'ocoa Sand 34 bulletin: museum of comparative zoology The time interval to be considered here is that between the deposi- tion of the Ocala limestone of Upper Eocene age and the withdrawal of the Hawthorne sea. During this time Florida alternated between an island and a peninsular condition several times. This unstableness appears to have been due as much to minor crustal movements as to major variations in sea level. No attempt has been made to correlate the crustal movements of this area with those of the Caribbean. All available data indicate that Florida was a submerged plateau throughout the Eocene. The close of the Eocene was marked by the beginning of a series of crustal movements which resulted in the formation of the Central Florida Dome. The distribution of the Oligocene and Miocene sediments indicate that the dome is the result of two sets of pressures working at nearly right angles to each other and at different times. That initiated at the close of the Eocene was a northwest-southeast pressure which affected the whole of the plateau and forced all but the southern portion out of water. At the same time there was a compensatory downwarping of the strata across the northern end of the plateau to form the Okefenokee Trough. At dif- ferent times the sea invaded this trough and cut Florida off from the mainland. During the Lower Oligocene, while the Marianna limestone was being deposited, the sea invaded the Okefenokee Trough only far enough to form a large bay at either end (Text fig. 5). The evidence for the bay in southeastern Georgia is the unusual thickness of the Oligocene sediments reported by Pretty man and Cave (1923) as being encountered in deep wells. However these sediments were identified mainly on lithological grounds. There was a renewal of the crustal movements at the close of Marianna time, which resulted in the formation of the Hatchatigbee Arch in western Alabama and in deepening of the Okefenokee Trough so that the Gulf communicated with the Atlantic across North Florida and South Georgia (Text fig. 6) through a strait, probably not over 50 or 60 miles wide. At this time the Glendon limestone was deposited. This period of deposition appears to have been brought to a close by a period of general emergence without any noticeable deformation of the land mass, and Florida was again connected with the mainland. The late Upper Oligocene was a period of general submergence which reduced Florida to a relatively small island in what is now the northwestern part of the peninsula (Text fig. 7). The strait which separated Florida from the mainland must have been nearly 150 white: lower MIOCENE MAMMAL FAUNA OF FLORIDA 35 miles wide. During this time the Flint River and Suwannee limestones were laid down. This period of deposition was brought to a close by a renewal of the crustal movements which marked the opening of the STRUCTURE MAP OF FLORIDA AND ADJACENT AREAS CONTOURS ON TOP OF OCA LA LIMESTONE Fig. 4. Structure map of Florida and adjacent areas from Mossom (1926), Cooke and Mossom (1928), and Prettyman and Cave (1923). Oligocene. At this time the Chattahoochie Arch in southwestern Georgia and southeastern Alabama was formed. The distribution of the Flint River sediments indicate that the main axis of the arch has a northeast-southwest direction rather than a north-south one as 36 bulletin: museum of comparative zoology postulated by Stephenson and Veatch (1915, p. 58). The formation of the arch seriously restricted the width of the trough in this area. During this same time interval the Florida Plateau was subjected to east-west pressure which shifted the long axis of the dome from an east-west direction to a north-south one. Stephenson and Veatch (1915, p. 59) applied the name Withlacoochie Anticline to the northern part of the Central Florida Dome. I have used the latter term because it seemed to convey a cleared concept of the doming of the Ocala limestone. Although the contact between the Suwannee and Tampa limestones has not been observed, the very different distributional patterns of the outcrops of the two limestones permit us to postulate an interval of uplift and a land connection between the island and the mainland during the interval between their depositions. Prettyman and Cave (1923, p. 82) report that the Chattahoochie ( = Tampa) limestone is conglomeratic in the southwestern part of Georgia and that it lies unconf ormably above the Glendon ( = Flint River) limestone. The very thin layer of limestone in the Ocala area containing Late Tampa Fossils (Mansfield, 1937, p. 24) indicate that there was con- tinuous subsidence of the island till the end of Tampa time. The thickness of the Tampa limestone reported in the deep wells which have been studied substantiate this. The thicknesses are: Monroe Co.— 300 ft. (Cole, 1941, p. 10); Polk Co. 140 ft. (ibid, p. 5); Gulf Co. 84 ft. (Cole, 1938, p. 9). Mansfield (1937, p. 31) reported a thickness of 89 ft. for the Tampa limestone in Gadsen Co. (Type locality of the Chattahoochie formation). Small discrepancies are inevitable in well records unless a continuous log is kept, but the differences shown here are much too great to attribute to normal error. Consequently it seems reasonable to assume that the younger zones of the Tampa limestone overlap the older. Concerning the interval between the Tampa and Hawthorne, Mans- field (1937, p. 44) says: "The lithological and the faunal difference between the Tampa limestone and the succeeding Alum Bluff group seem too great to attribute solely to shifting of shore-line currents and therefore strongly suggest a period of uplift at the end of Tampa time, followed by subsidence and deposition of the Alum Bluff group." Cushman and Ponton (1932, p. 31) say: "However, in the case of the Chipola formation along the Chipola River, at and near its type locality, we find the soft, greenish-gray shell marl of the Chipola lying on what looks like the eroded surface of the hard, white to buff lime- stone of the Tampa." This area appears to be rather close to the north white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 37 boundary of the Okefenokee Trough and could be only a local uncon- formity. However at Rock Bluff in Liberty Co. the section is very different. Cooke and Mossom (1928, p. 119) say: "It is difficult to SURFACE AND SUBSURFACE DISTRIBUTION OF THE MAR/ANNA LIMESTONE Fig. 5. Distribution of Marianna limestone (stippled portion). draw the line precisely between the Tampa limestone and the Haw- thorne formation at Rock Bluff, for one seems to grade perfectly into the other." Rock Bluff is about 20 miles north of what is believed to have been the middle of the Okefenokee Trough. Consequently there is no good evidence that the trough was drained at this time. 38 bulletin: museum of comparative zoology The distribution of the outcrops of the Tampa limestone and Haw- thorne formation show that the western side of the Florida Plateau was elevated at the close of Tampa time so that the bulk of the land SURFACE AND SUBSURFACE DISTRIBUTION OF THE GL ENDON L IMES TONE Fig. 6. Distribution of Glendon limestone. lay west of the axis of the plateau. In the excavations in Gilchrist Co., a boulder bar (some of the boulders were fifteen inches in diameter) was encountered in the old stream channel in which the fossils are found. It seems logical that this bar should have been formed at this interval of uplift. The elevation of the west coast alone would have white: lower miocene mammal fauna of Florida 39 given the stream sufficient gradient to transport the boulders. There is no evidence in the fossil mammals for postulating a land bridge be- tween the island and the continent at this time. SURFACE AND SUBSURFACE DISTRIBUTION OF THE FLINT RIVER AND EQUIVALENT FORMATIONS Fig. 7. Distribution of Flint River and equivalent formations. Because of the very poor state of preservation of the fossils found in the Hawthorne formation, there is considerable doubt concerning the amount of time that it represents. Cooke and Mossom (1928, p. 98) say: "In the Peninsula the Alum Bluff group is represented by the Hawthorne formation. Fossils obtained from the Hawthorne show 40 bulletin: museum of comparative zoology that at least part of it is of Chipola age, but its fauna at some localities seems to be younger than the Chipola." On page 110 of the same pub- lication they report that the Oak Grove fauna is found on hilltops near SURFACE AND SUBSURFACE DISTRIBUTION OF FHE TAMPA LIMESTONE Fig. 8. Distribution of Tampa limestone. Bainbridge, Georgia and Roberts, Escambia Co., Alabama. Cooke (1935, p. 100) assigns the Alum Bluff of Georgia to the Hawthorne formation. It would appear then that the Hawthorne included beds equivalent in age to the Oak Grove Sand, but, as far as I have been able to learn, the Shoal River fauna has not been recognized in the Hawthorne formation. white: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 41 At Alum Bluff, 4j/£ miles north of Blountstown, Liberty Co., Florida Cooke and Mossom (1928, p. 108) record an unconformity between the Chipola formation and the overlying, plant-bearing beds whose age is / THOMAS FARM 2 MIDWAY 3 GRISCOM PLANTATION SURFACE AND SUB5URFACE DISTRIBUTION OF THE HAWTHORNE AND EQUIVALENT FORMATIONS Fig. 9. Distribution of Hawthorne and equivalent formations. still in doubt. This locality is so close to the north boundary of the Oke- fenokee Trough that the unconformity could be caused by a fluctuation of the shore line that would not necessarily drain the trough. As far as I have been able to learn this is the only unconformity recorded in the Hawthorne formation. 42 bulletin: museum of comparative zoology Cushman and Ponton (1932) seem to be of the opinion that there was no wide spread interruption in deposition during Alum Bluff time. They (1932, p. 32) say: "The shore-line during this period was a very variable feature, advancing and retreating, depositing beds of varying thickness but all comparatively thin, and forming over-laps to such an extent that it is possible that such over-laps might be mistaken for an unconformity, even between successive zones at any one exposure." If the unconformity above the Chipola formation at Alum Bluff is only a local feature, the mammalian fauna of the island would have no opportunity to communicate with that of the mainland till after the close of Oak Grove time. This agrees very well with the present corre- lation of the continental Miocene of the Great Plains. Tomarctus thomasi, Amphicyon longiramus, Parahippus barbouri, P. leonensis, and Merychippus gunteri of the Lower Miocene fauna of Florida are very closely related to Tomarctus optatus and brccirostris, Amphicyon sinapius, Parahippus coloradcnsis and brevidcns, P. cognatns and brevidens, and Merychippus primus respectively, of the Late Middle and Upper Miocene of the Plains. That the latter group have no close relatives in the Lower Miocene fauna of the Plains lends support to the idea that they descended from the Florida stock, but were unable to reach the Plains till after Oak Grove time. It seems probable that the mammalian fossils found at Quincy, Mid- way, and Griscom Plantation (since they occur in marine sediments) had been washed out to sea after severe storms either as carcasses or as individual bones. ENVIRONMENT During the time period represented by the fluvial deposit in Gilchrist Co., Florida was a limestone island cut off from the mainland by a shallow sea fifty or sixty miles wide. In Tampa time Florida was an elliptical island roughly 220 miles north-south by 100 east-west. The crustal movements at the end of Tampa time shifted the shore line some but only slightly increased the width. The strait which separated the island from the continent was not appreciably wider during Haw- thorne time than during the Tampa. If the structure of Florida (Text fig. 4) during the Lower Mio- cene was at all similar to that of today the highest part of the island would have had an elevation of about 200 feet. This is not enough seriously to affect the climate. There is no reason to suppose that the climate was verv different then than now. white: lower miocene mammal fauna ok Florida 43 The bedrock of the central part of the island was formed by the Ocala limestone and the Suwannee limestone formed that around the edges. Both of these limestones are granular and poorly consolidated. Also both are very pure, but of the two the Suwannee limestone carries a slightly higher content of impurities. Concerning the Ocala lime- stone Cooke and Mossom (1928, p. 48) say: "Its texture is commonly granidar, but parts of it have been converted to hard, compact rock by the deposition of travertine or calcite in its interspaces. In some places it consists of a loosely coherent mass of Foraminifera, Bryozoa, and other small organisms, a mass so porous that water can percolate freely through it; elsewhere it is finer grained and more compact, al- though still pervious to water. "In chemical composition, as in physical character, the Ocala lime- stone is remarkably uniform. It consists almost entirely of carbonate of lime, and in places contains as little as four-tenths of one percent of impurities." The same statements are essentially true of the Suwannee limestone. Of it Mansfield (1937, p. 46) says: "The formation consists almost en- tirely of limestone. The unweathered rock is granular to dense, com- pact, usually cream-colored, rather pure limestone. The lower part is at many places more granular than the upper. Mossom gives the following analysis of Suwannee limestone for the quarry of the Florida Hard Rock Products Co., Brooksville, Fla.: "Silica (Si02) (3.54 Iron and alumina (Fe+Al) 1.44 Calcium carbonate (CaC03) 91.09 Magnesium carbonate (MgC03) trace Undetermined .93 100.00" There is, in the Tampa area, a bed of "tough, plastic, greenish sandy clay" 41 to 64 feet thick (Mansfield, 1937, p. 14) below the Tampa limestone. This bed has been encountered in wells and is not known as an outcrop. A somewhat similar deposit is recorded (Mansfield, ibid, p. 29) below the Tampa limestone at Wyley Landing, Georgia. Similar deposits of this age are as yet unknown elsewhere in Florida. It is possible that both are delta deposits and very local in extent. From the above it is evident that the soil of this island, during the Lower Miocene, was a soft, porous, very pure limestone into which the plants could force their roots without much difficulty. Also these 44 bulletin: museum of comparative zoology limestones are highly fossiliferous and the hard parts of marine organ- isms are in sufficient abundance to supply more than the necessary amount of phosphorus. The abundance of sinkholes in Florida today is ample evidence that these limestones were sufficiently soluble to be available to plants as food. Undoubtedly the soil conditions were similar to that in the Miami-Homestead area today. This area sup- ports abundant vegetation when it receives sufficient rainfall. Unfortunately there are no fossil plants found in the deposit in which the mammalian remains occur. However, it is permissible to suppose that the vegetation was similar to that found in the Miami- Homestead area today. It is interesting that cement on the teeth of horses in quantity to be functionally advantageous should make its geologically earliest appear- ance in a region whose soil contained a superabundance of calcium and phosphorus. The deposition of cement around the roots of the teeth and a small amount at the base of the crown is nearly universal in the mammals. That the cement depositing organ should suddenly begin to deposit an excess of cement around the teeth is obviously the result of a change in the animal's physiology. The progressive specimens were able to utilize some of the abundant calcium, or what is just as probable and just as important, they may have lacked the ability to excrete the excess calcium and were forced to utilize it. The deposition of cement on the crown of the teeth was a genetically unstable character in all of the species of fossil horses in this deposit except Merychippus gun- teri. In this species it arrived at a more or less stable quantity after a series of progressive changes from species to species, each having more than its parent, but each presenting a rather wide variation within the species. It would seem then that the deposition of a large amount of cement on the teeth was a hereditary factor which gained complete dominance in this stock at the time it reached the Merychippus stage of development. CONCLUSIONS The lenticular nature of the laminated clays and the unsorted ma- terial in the boulder bar show that the mammal bearing sediments were deposited in the channel of a stream which had arrived at grade. Although the data are not available to date exactly the beginning and end of the time interval represented by this deposit it is evident that it began after the deposition of the Suwannee limestone and was brought to a close by the invasion of the Hawthorne sea. The time interval is undoubtedly equivalent to most of the Tampa limestone WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 45 and the lower part of the Hawthorne formation. The Carnivora and the Artiodaetyla have a decided Lower Miocene aspect, and both groups contain at least one genus which is not found on the Great Plains in deposits later than the Upper John Day. The Equidae and some of the Carnivora are very progressive and have as their closest relatives, forms which are found in the Middle and Upper Miocene deposits of the Plains. This disagreement with the stratigraphical data has a plausible explanation in that these forms developed on insular Florida and were unable to escape to the mainland till the end of Oak Grove time. It is thought (on the basis of the difference in the distribution of the Flint River and Tampa sediments) that the conti- nental fauna had an opportunity to reach Florida during the Flint River-Tampa interval, but the evidence is not conclusive. If this should eventually prove to be untrue and that the only opportunity for this fauna to reach Florida was during the Glendon-Flint River interval it will be necessary to revise downward the correlation of the Upper Oligocene and Basal Miocene deposits of the Great Plains. In only one species (Parahippus leonensis Sellards) are there enough specimens to note the amount of variation. It is possible to divide this series into three groups on the basis of the extremes of the variations which I shall refer to as the conservatives, progressives, and aberrant. The intergrades are sufficiently numerous that no specific separation is possible. The aberrations are not of a lethal or deleterious nature and show up occasionally in the milk teeth of Mcrychippus. They usually express themselves in unusual patterns of the crochet and as multiple plications on the metaloph. This may be the result of unusual activity of the endocrine glands during the time of the forma- tion of the permanent teeth. In the case of the conservative and pro- gressive specimens the difference may be that between sexes. If the principle of Unequal Growth is the correct explanation of the process of the folding of the enamel on the metaloph, it seems reasonable to suppose that the females, with a lower rate of metabolism, would have a less rapid proliferation of the enamel depositing cells and a simpler enamel pattern. However this question cannot be answered till asso- ciated skeletons are found. This series of fossil horses furnish relatively detailed data on specia- tion in this group. Although the series is not complete, the gaps are so small that they can be bridged with no difficulty. These species {Parahippus blackbergi-barbouri-leonensis-Merychippus gunteri) are a monophyletic series illustrating the development of Merychippus from a very primitive species of Parahippus. In the end there is a condition 46 bulletin: museum of comparative zoology in which the difference between these two genera is no greater than that which exists between two species of the same genus. Speciation Mtohi, LPP us rarah ipp< U3 Merychtppui 5: Niobrara River Lower Snaki Creek Cognatt us ^Pa Mascall Deep River vwnee Creek crentdens Coloraaens'is breviaens se/unctus 1 oheep Creek Garvin Gully Ce oar Run blackbergi Oak Grove ilL yea 1 cans R nmus Hawthorne Tampa I •0 I [2 leorjt ensis barbout blackbergi. aunt en Flint Rt ver I am pa Interval Fig. 10. Graphic representation of the relationships of some of the species of fossil horses. proceeded rather rapidly and the whole process took place in a rather short period of time. Nor was the entire population transformed from one species into another but the new species split off from the parent WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 47 species by series of progressive changes and the two existed side by side for most of the remainder of the Miocene. Soon after the new species split oft' from the parent species it gave rise to another new species and so on. Unfortunately the record stops with the production of Merychippus gunteri. It would seem that a "high biotic potential" appeared in this stock and was passed on to each new species by robbing the parent species, although the parent species persisted, with di- minishing abundance, till near the close of the Miocene. The idea is graphically represented in Text fig. 10. This is entirely compatible with the record of the appearance of major groups in geologic time. The data on speciation obtained from this series indicate that the phylo- genetic relationships of species based on a series from a single time unit are more nearly correct than formerly supposed, because this pic- ture would be essentially the same whether the series was taken from the Hawthorne, or Sheep Creek, or Deep River. It is obvious from the above data that Natural Selection could have played no part in the development of Meri/chippus and that the causes of speciation are to be found within the animals themselves. The identification of the causes is the work of the physiologist, the endocrinologist, and the biochemist. The paleontologist can only catalogue the changes after they have taken place. 48 bulletin: museum of comparative zoology Geological Literature Campbell, R. B. 1940. Outline of the Geological History of Florida. Proc. Fla. Acad. Sri., 4, p. 87. Cole, W. S. 1938. Stratigraphy and Micropaleontology of Two Deep Wells in Florida. Florida Dept. Conserv., Geol. Bull., No. 16. 1941. Stratigraphic and Paleontologic Studies of Wells in Florida. Fla. Dept. Conserv., Geol. Bull., No. 19. Cooke, C. W. 1923. The Correlation of the Vicksburg Group. U.S.G.S. Prof. Paper 133. 1926. The Cenozoic Formations in the Geology of Alabama. Geol. Sur- vey of Alabama, Special Report No. 14. 1935. Notes on the Vicksburg Group. Am. Ass. Petrol. Geol., Bull., 19. No. 8, p. 1162. 1936. Geology of the Coastal Plain of South Carolina. U.S.G.S., Bull. 867. 1939. The Boundary between the Oligocene and Miocene. Am. Ass. Pet" rol. Geol., Bull., 23, p. 1560. 1940. Prefatory Note in W. C. Mansfield's Mollusks of the Chickasawhay Marl. Journ. Paleont., 14, p. 171. Cooke, C. W. and Mansfield, W. C. 1936. Suwannee Limestone of Florida. Proc. Geol. Soc. Amer. for 1935, p. 71. Cooke, C. W. and Mossom, Stuart 1928. Geology of Florida. Florida State Geol. Survey, Twentieth Ann. Report. Cushman, J. A. and Ponton, G. M. 1932. The Foraminifera of the Upper, Middle and part of the Lower Miocene of Florida. Florida State Geol. Survey, Bull. No. 9. Howe, Henry V. 1942. Fauna of the Glendon Limestone at its Type Locality. Journ. Paleont., 16, p. 264. Mansfield, W. C. 1937. Mollusks of the Tampa and Suwannee Limestones of Florida. Fla. Dept. Conserv., Geol. Bull. No. 15. WHITE: LOWER MIOCENE MAMMAL FAUNA OF FLORIDA 49 Matson, G. C. and Sanford, Samuel 1913. Geology and Ground Waters of Florida. U.S.G.S. Water-supply paper 319. Mossom, Stuart 1926. A Review of the Structure and Stratigraphy of Florida with Special Reference to the Petroleum Possibilities. Florida State Geol. Survey, 17th Ann. Report, p. 169. Prettyman T. M. and Cave, H. S. 1923. Petroleum and Natural Gas Possibilities in Georgia. Geol. Survey of Georgia, Bull. No. 40. Semmes, D. R. 1929. Oil and Gas in Alabama. Geol. Survey of Alabama, Special Re- port 15. Stephenson, L. W. and Veatch, J. O. 1915. Underground Waters of the Coastal Plain of Georgia. U.S.G.S. Water-supply Paper 341. Toulmin, Lyman D., Jr. 1940. The Salt Mountain Limestone of Alabama. Geol. Survey of Alabama, Bull. 46. Veatch, J. O. and Stephenson, L. W. 1911. Geology of the Coastal Plain of Georgia. Geol. Survey of Georgia, Bull. 26. EXPLANATION OF PLATES PLATE 1 White — Lower Miocene Mammal Fauna of Florida PLATE 1 Fig. 1. Daphaenus caroniavorus spec. nov. Type, M.C.Z. No. 3727, Crown view of left M1^. x 2%. Fig. 2. Paradaphaenus tropicalis spec. nov. Type, M.C.Z. No. 3729, right PMVR x2H- Fig. 3. Nothocyon insularis spec. nov. Type, M.C.Z. No. 3812, occlusal view of right P4-M2. x 2M. BULL. MUS. COMP. ZOOL White Lower Miocene Mammal Fauna of Florida. Plate 1 M.C.Z. 372.7 V rW- MC.Z. 3729 M.C.Z. 3812 PLATE 2 White — Lower Miocene Mammal Fauna of Florida PLATE 2 Fig. 1. Paradaphaenus nobilis (Simpson). M.C.Z. No. 3725, occlusal view of left P"-M2. x 23^. Fig. 2. Tomarctus canavus (Simpson). M.C.Z. No. 3813, occlusal view of leftPMVP. x2j^. BULL. MUS. COMP. ZOOL White: Lower Miocene Mammal Fauna of Florida. Plate2 PLATE 3 White — Lower Miocene Mammal Fauna of Florida PLATE 3 Fig. 1. Lateral and Fig. 2, occlusal views of Paradaphaenas nobilis (Simp- son). M.C.Z. No. 3724. x 1. BULL. MUS. COMP. ZOOL. White Lower Miocene Mammal Fauna of Florida. Plate 3 PLATE 4 White — Lower Miocene Mammal Fauna of Florida PLATE 4 Fig. 1. Occlusal, and Fig. 2, lateral views of Paradaphaenus tropicalis spec. nov. Paratype, M.C.Z. No. 3714. x 1. BULL. MUS. COMP. ZOOL. White: Lower M iocene Mammal Fauna of Florida. Plate 4 PLATE 5 White — Lower Miocene Mammal Fauna of Florida PLATE 5 Fig. 1. Occlusal, and Fig. 2, lateral views of Amphicyon longiramus spec, nov. Type, M.C.Z. No. 3919. x 1/2. BULL. MUS. COMP. ZOOL. White Lower Miocene Mammal Fauna of Florida. Plate 5 PLATE 6 White — Lower Miocene Mammal Fauna of Florida PLATE 6 Fig. 1. Medial, Fig. 2, occlusal, and Fig. 3, lateral views of Tomarctus canavus (Simpson). M.C.Z. No. 3628. x 1. BULL. MUS. COMP. ZOOL. White: Lower Miocene Mammal Fauna of Florida. Plate 6 Fig. 2 M.CZ.3628 PLATE 7 White — Lower Miocene Mammal Fauna of Florida PLATE 7 Fig. 1. Occlusal, and Fig. 2, lateral views of Tomarctus thomasi White. M.C.Z. No. 3712. x 1. BULL. MUS. COMP. ZOOL. White: Lower Miocene Mammal Fauna of Florida. Plate 7 sm m* <;f v PLATE 8 White — Lower Miocene Mammal Fauna of Florida PLATE 8 Fig. 1. Palatal, and Fig. 2, lateral views of skull of Hypcrmekops olseni gen. et spec. nov. Genoholotype, M.C.Z. No. 3711. x 4/10. Fig. 3. Occlusal view of left P2-M3. x 4/5. BULL. MUS. COMP. ZOOL. White; Lower Miocene Mammal Fauna of Florida. Plate 8 PLATE 9 White — Lower Miocene Mammal Fauna of Florida PLATE 9 Fig. 1. Lateral, and Fig. 2, occlusal views of left mandible of Aiichitherium clarencei Simpson. M.C.Z. No. 3810. x 2/3. BULL. MUS. COMP. ZOOL. White: Lower Miocene Mammal Fauna of Florida. Plate9 PLATE 10 White — Lower Miocene Mammal Fauna of Florida PLATE 10 Fig. 1. Occlusal view of left upper dentition of Parahippus blackbergi (Hay). M.C.Z. No. 3829. x iy2. Fig. 2. Lateral, and Fig. 3, occlusal view of milk teeth and permanent molars of Parahippus blackbergi (Hay). M3 has been artificially exposed. A/I ri V -NT,, oo An -il/ M.C.Z. No. 3840. x \V2. BULL. MUS. COMP. ZOOL. White Lower Miocene Mammal Fauna of Florida. Plate 10 >/^ WHi %m ^^ .a '■^imi m PLATE 11 White — Lower Miocene Mammal Fauna of Florida PLATE 11 Fig. 1. Occlusal, and Fig. 2, lateral views of upper right cheek teeth of Parahippus barbouri spec. nov. Type, M.C.Z. No. 3646. x 1. BULL. MUS. COMP. ZOOL. White Lower Miocene Mammal Fauna of Florida. Plate 11 V\ PLATE 12 White — Lower Miocene Mammal Fauna of Florida PLATE 12 Fig. 1. Occlusal, and Fig. 2, lateral views of right mandible of Parahippus barbouri spec. nov. Paratype, M.C.Z. No. 3814. x 1/2. BULL. MUS. COMP. ZOOL. White: LowerMiocene Mammal Fauna of Florida. Plate 12 PLATE 13 White — Lower Miocene Mammal Fauna of Florida PLATE 13 Fig. 1. Conservative (M.C.Z. No. 3921), and Fig. 2, progressive (M.C.Z. No. 3744) specimens of Parahippus leonensis Sellards. xl|. Fig. 3. Merychippus gunteri Simpson. M.C.Z. No. 3801. Occlusal view of upper right cheek teeth, x 4/5. BULL. MUS. COMP. ZOOL. White; Lower Miocene Mammal Fauna of Florida. Plate 13 -/^ 'W7 3-TlH s N PLATE 14 White — Lower Miocene Mammal Fauna of Florida PLATE 14 Details of the Excavation in North Florida. Fig. 1, diagrammatic plan of excavation. The dotted line indicates the approximate limits of the boulder bar, and the dashed line indicates the edge of the outlier of the Hawthorne formation. Fig. 2, Generalized section along line BJ-B. Fig. 3, Generalized section along line C'-C. Fig. 4, generalized section along line AJ-A. BULL. MUS. COMP. ZOOL. White: Lower Miocene Mammal Fauna of Florida. Plate 14 Hi ir, CM -J it II So N ^ ^1 in 1 1 I ^ 0 •1 I'll m a -* 3-hr 0> * Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE Vol. XCII, No. 2 FIRST SUPPLEMENT TO THE LIST OF TYPES OF BIRDS NOW IN THE MUSEUM OF COMPARATIVE ZOOLOGY By James L. Peters CAMBRIDGE, MASS., U.S.A. PRINTED FOR THE MUSEUM March, 1943 No. 2. — First Supplement to the List of Types of Birds now in the Museum of Comparative Zoology By James L. Peters In March 1930 was published in this Bulletin (70, 4, p. 147-426) a list of the types of birds then in the collection. The author was Outram Bangs, Curator of Birds, who between 1909 and the time of his death in 1932 built up the museum's bird collection from a small and inadequate one to one of the finest anywhere. The List of Types was published as an anniversary volume and was first distributed on the evening of 17 March, 1930 at a meeting of the Nuttall Ornithological Club, when Bangs' fiftieth anniversary of election to membership in the Club was appropriately celebrated. Since the publication of the first list of types, an open manuscript has been kept in anticipation of publishing a supplement at some future date. Comments on types made by Outram Bangs in the open manuscript are initialed O. B. Otherwise the author is responsible for such discussion. The time for publication of the supplemental list has now arrived, since it seems probable that no important ornitho- logical discoveries based on exploration and new field work will be made for some years to come. The order of this supplemental list is the same as that employed in the first list, i.e., that of Sharpe's Hand-List. The dagger (f ) indicates that the name is surely a synonym. This also seems an opportune time to publish a bibliography of the writings of Outram Bangs ; no such list has been published previously, and since Bangs' period of activity extended for nearly forty years and his articles appeared in many different journals, it is doubtful if a complete bibliography could be prepared without access to his own carefully kept set. The task of preparing this bibliography was entrusted to Miss Margaret D. Porter (now Mrs. Chandler Bigelow) formerly a Research Assistant in the Bird Department, whose catalogues and card entries bear testimony to the many hours of painstaking care she put into her work. 54 bulletin: museum of comparative zoology TINAMIDAE Crypturellus soui decolor Griscom and Greenway Crypturellus soui decolor Griscom and Greenway, Bull. Mas. Comp. Zool., 81, 2, May ( = 10 June), 1937, p. 417. Type. No. 173012, ad. c?; Brazil: Para; Pinhy, on the right bank of the Rio Tapajoz; 15 June, 1933; A. M. Olalla. Listed as No. 173021 in the original description due to typographical error. CRACIDAE Penelope superctliaris argyromitra Neumann Penelope superciliaris argyromitra Neumann, Bull. Brit. Orn. Club, 53, 31 Jan- uary, 1933, p. 94. Type. No. 160966, ad. 9 ; Brazil: Central Goyaz, Veadeiros, near Forte, "2/12/" 1929; Jose Blaser. Compared with birds from the Tapajoz and Santarem, Neumann's type differs chiefly in that the whitish superciliaries meet across the forehead; the superciliaries are broader than in specimens from Bahia and Sao Paulo. The type is much grayer (less green or bronzy) than any specimens of P. superciliaris in the Museum of Comparative Zoology; it is also in much more worn plumage than any skins of the species we possess. Forte is in east-central Goyaz on the Rio Paranan, about 150 miles from its junction with the Tocantins. Ortalis wagleri griseiceps van Rossem Ort-alis wagleri cinereiceps van Rossem, Bull. Mus. Comp. Zool., 77, 7, 29 December, 1934, p. 431. Type. No. 224937, ad. ; Philippine Islands: Basilan, 15 km. northeast of Maluso; 23 April, 1937; Barbara Lawrence. Rhipidura teijsmanni sulaensis Neumann Rhipidura teijsmanni sulaensis Neumann, Bull. Brit. Orn. Club, 59, 21 April, 1939, p. 93. Type. No. 269600, ad.cT; Sula Islands: Taliabu, 11 October, 1938; J. J. Menden. Myiagra azureicapilla azureicapilla Layard Myiagra azureicapilla Layard, Ibis, 1875, p. 434. Cotype. No. 166780, ad.cT; Fiji Islands: Taviuni, Ngila; 18 August, 1875; E. L. Layard. Cotype. No. 166781, ad. 9 ; Fiji Islands: Taviuni, Ngila; 11 August, 1875; E. L. Layard. This species was described by Layard who neither designated a type, nor stated the number of specimens that he had. The 4th volume of the Catalogue of Birds in the British Museum lists a male and female from Ngila, Taviuni, but does not claim either specimen as a type. A pair also went to Rowley who figured them in his "Orni- thological Miscellany," 1, pi. 35. The two specimens which I claim as cotypes are the two figured birds, bought at the Rowley auction in November, 1934 by Rosenberg, the London dealer, and from whom the museum obtained them. Mathews makes this species the type of his monotypic genus Lophomyiagra, a genus that may have to be recognized eventually, but pending a general review of the Muscicapidae I make no change. peters: supplementary list of types of birds 81 CAMPEPHAGIDAE Edolisoma morio talautense Meyer and Wiglesworth Edoliisoma talautense Meyer and Wiglesworth, Abh. Ber. K. Zool. Mus. Dresden, 1894-95 (1895), no. 9, p. 5. Cotype. No. 97336, ad.cf; Talaut Islands: Esang; 20 October, 1894; Charles W. Cursham's collectors. Received in exchange with the Dresden Mu- seum. (No. C 13800). Meyer and Wiglesworth described this form from fifteen specimens from the islands of Karkellang, Esang and Kabruang in the Talaut group; no holotype was designated. In their Birds of Celebes (2, 1898, p. 423) they list "ad. 9 , type, Karkellang, Nov. 1894: Nat. Coll. — C 13795 and others" and "ad.cf, type, Kabruang, Nov. 1893 — -C 13121, and others." As in the case of Tanygnathus talautensis, Meyer and Wiglesworth's action in the Birds of Celebes amounts to a subsequent selection of a male and female cotype, but I cannot see how it invalidates the right of any of the other original specimens to rank as cotypes. Pericrocotus miniatus dammermani Neumann Pericrocotus miniatus dammermani Neumann, Bull. Brit. Orn. Club, 57, 30 June, 1937, p. 152. Type. No. 177810, 9 ; South Sumatra: Gunong Dempo, 2500 metres; 20 July, 1936; J. J. Menden. No specimen identified by a museum number was designated in Professor Neumann's original description as the type; but his state- ment in the introduction to the description of this and three other subspecies of birds that the types are in the Museum of Comparative Zoology, coupled with the fact that the specimen is marked "typus" in its describer's handwTiting and that the data correspond, definitely establishes its right to be the holotype. TIMALIIDAE Garrulax moniliger schauenseei Delacour and Greenway Garrulax moniliger schauenseei Delacour and Greenway, Bull. Brit. Orn. Club, 59, 17 June, 1939, p. 132. Type. No. 265100, d"; Laos: Xieng-Khouang, 1200 metres; 6 November (in original description), 6 December (on label), 1938; J. Delacour, J. C. Greenway, Jr., F. Edmond-Blanc. Field no. 76, VII Exped. en Indo- Chine. 82 bulletin: museum of comparative zoology Rhinocichla mitrata griswoldi Peters Rhinocichla mitrata griswoldi Peters, Bull. Mus. Comp. Zool., 87, 3, December, 1940, p. 204. Type. No. 236020, ad., not sexed; Borneo: Mt. Tibang, 4000 feet; 19 Novem- ber, 1925; Eric Mjoberg. This specimen while not sexed is doubtless a male since it has a wing measurement of 109 mm., about the maximum for that sex. Wings of females run from 95.5 to 104.5. Neocichla gutturalis angustus Friedmann Neocichla gutturalis angustus Friedmann, Journ. Wash. Acad. Sci., 20, 17, 1930, p. 434. Type. No. 134447, ad. 9 ; Tanganyika Territory: Muhalala, Kilamatindi; 3 March, 1922; Arthur Loveridge. Turdinus rufipennis distans Friedmann now Illadopsis rufipennis distans (Friedmann) Turdinus rufipennis distans Friedmann, Proc. New England Zool. Club, 10, 14 April, 1928, p. 48. Type. No. 237750, cT1; Tanganyika Territory: Amani, Usambara Mts.; 22 November, 1926; A. Loveridge. The type was in the Museum of Comparative Zoology at the time that Outram Bangs published the list of types, but was accidentally omitted from his list. This form is very distinct from the typical race; Friedmann dis- cussed the differences between distans and allied forms with fairly good material available and his conclusions as to its relationship may be accepted. Illadopsis stictigula pressa Bangs and Loveridge Illadopsis stictigula pressa Bangs and Loveridge, Proc. New England Zool. Club, 12, 1931, p. 94. Type. No. 148499, ad. 9 ; Tanganyika Territory: Nkuka Forest, Rungwe Mountains; 5 April, 1930; Arthur Loveridge. Napothera epilepidota mendeni Neumann Napothera epilepidota mendeni Neumann, Bull. Brit. Orn. Club, 57, 30 June, 1937, p. 152. Type. No. 177863, 9 ; South Sumatra: Gunong Dempo, 1800 metres; 21 July, 1936; J. J. Menden. peters: supplementary list of types of birds 83 No specimen identified by a museum number was designated as the type in Professor Neumann's original description, but his statement in the introduction to the paper in which this and three other sub- species of birds were named that the types are in the Museum of Comparative Zoology coupled with the fact that this was the only specimen of this form received and that the label is marked in Neu- mann's own hand "Typus von Napothera cpilepidoia mendeni Neum.", definitely establish this specimen as the holotype. Alcippe ruficapilla danisi Delacour and Greenway Alcippe (Fulretta) ruficapilla da nisi Delacour and Greenway, Proc. New England Zool. Club, 18, 3 May, 1941, p. 47. Type. No. 268092, ad. cf; Laos: Xieng-Khouang; 15 December, 1938: J. Delacour and J. C. Greenway, Jr. Stachyris striolata helenae Delacour and Greenway Stachyris -striolata helenae Delacour and Greenway, Bull. Brit. Orn. Club, 59, 17 July, 1934, p. 130. Type. No. 265102, estris aharonii Neumann, Anz. Orn. Ges. Bayern, 2, 8 March, 1934, p. 333. Type. No. 160990, ad.d"; Syria: Ras Baalbek; 17 April, 1931; T. Aharoni. FRINGILLIDAE Pipilo fuscus texanus van Rossem Pipilo fuscus texanus van Rossem, Trans. San Diego Soc. Nat. Hist., 7, 34, 31 May, 1934, p. 371. Type. No. 316022, (formerly 16025, Thayer Collection), ad.d"; Texas: Kerrville; 24 April, 1910; F. B. Armstrong. Colonel Thayer made a slight error in cataloguing six specimens of Pipilo fuscus all collected by Armstrong at Kerrville, Texas. Two of the Thayer birds bore his number 16025. The only specimen, however, which agrees with the date and sex of the type is listed in the catalogue under 16022, but none of the skins catalogued from 16020 and 16025 94 bulletin: museum of comparative zoology inclusive bear this number on their labels. The seventh specimen of the series was taken in Kerr County, Texas, 8 April, 1914, also by Arm- strong. Pipilo fuscus perpallidus van Rossem Pipilo fuscus perpallidus van Rossem, Bull. Mus. Comp. Zool., 77, 7, 29 December, 1934, p. 483. Type. No. 222952; Chihuahua: Chihuahua; 30 November, 1888; M. Abbott Frazar. Pipilo maculatus griseipygius van Rossem Pipilo maculatus griseipygius van Rossem, Bull. Mus. Comp. Zool., 77, 7, 29 December, 1934, p. 482. Type. No. 222899, ad. cf; Chihuahua: Jesus Maria; undated, but probably the autumn of 1884; R. R. McLeod. Pipilo maculatus consobrinus Ridgway Pipilo maculatus consobrinus Ridgway, Bull. Geol. and Geogr. Surv. Terr., 2, 2, April, 1876, p. 189. Cotijpe. No. 328590, ad. & ; Lower California: Guadelupe Island; 20 February, 1875; Dr. Edward Palmer. This specimen is one of the original series taken by Dr. Edward Palmer on which Ridgway based his description. Following the usual custom in such cases, I consider that all specimens of the type series should rank as cotypes. This bird was exchanged by the United States National Museum to Col. John E. Thayer several years ago; Col. Thayer had it mounted and placed on exhibition in his beautiful little private museum. Together with other mounted birds, this speci- men came to the Museum of Comparative Zoology after Col. Thayer's death. It has been relaxed and is once more in the form of a skin. Passerina leclancheri grandior Griscom Passerina leclancheri grandior Griscom, Bull. Mus. Comp. Zool., 75, 1934, p. 420. Type. No. 238393, ad.d" ; Oaxaca: Chivela; 19 March, 1927; W. W. Brown. Melozone rubricatum grisior van Rossem Melozone rubricatum grisior van Rossem, Trans. San Diego Soc. Nat. Hist., 7, 23, 31 March, 1933, p. 283. Type. No. 222695, ad. a*; Sonora: Hacienda de San Rafael; 11 May, 1888; M. Abbott Frazar. peters: supplementary list of types of birds 95 At the time that Frazar collected at Hacienda de San Rafael, it was located in extreme western Chihuahua. Since then the Sonora- Chihuahua boundary has been relocated and it appears that the Hacienda now lies within the borders of the State of Sonora as mapped today. The relocation of political boundaries, reapportionment of countries among the great powers and the renaming of territories, while of absolutely no zoographical significance, bring about much confusion in correctly locating type localities of fifty years ago by present day maps. Amphispiza bilineata confinis van Rossem Amphispiza bilineata confinis van Rossem, Bull. Mus. Comp. Zool., 77, 7, 29 December, 1934. Type. No. 222576, ad.cT; Chihuahua: Chihuahua; 12 November, 1888; M. Abbott Frazar. Xenospiza baileyi Bangs Xenospiza baileyi Bangs, Proc. New England Zool. Club, 12, 1931, p. 87. Type. No. 45986, ad.c?; Jalisco: Bolaiios; 8 March, 1889; (W. B. Richard- son?). I know of no instance that could possibly give better evidence of Outram Bangs' retentive memory for every bird skin that he ever handled than the circumstances that led to the naming of this bird. As he explained in the original description, the skin of the type had lain unnamed for many years in a "first series" tray. When he unpacked a finch that Alfred M. Bailey had recently collected in Mexico and forwarded to the M. C. Z. for identification, without saying a word Bangs went directly to the tray in question, selected the skin whose identity had been a mystery for so many years, compared it with Bailey's freshly collected specimen and lo, the two matched. Aimophila quinquestriata septentrionalis van Rossem Aimophila quinquestriata septentrionalis van Rossem, Bull. Mus. Comp. Zool., 77, 7, 29 December, 1934, p. 485. Type. No. 222625, ad. cT; "Chihuahua" {i.e. Sonora): Hacienda de San Rafael; 18 May, 1888; M. Abbott Frazar. Aimophila humeralis asticta Griscom Aimophila humeralis asticta Griscom, Bull. Mus. Comp. Zool., 75, 1934, p. 417. Type. No. 111800, ad.d"; Colima: Colima; 20 January, 1889; W. B. Richard- son. 96 bulletin: museum of comparative zoology AlMOPHILA RUFESCENS ANTONENSIS Van Rossem Aimophila rufescens antonensis van Rossem, Trans. San Diego Soc. Nat. Hist., 9, 36, 1 October, 1942, p. 436. Type. No. 114601, ad. cf ; Sonora: La Chumata Mine, 4500 feet, Sierra de San Antonio; 23 May 1905; W. W. Brown. The type locality is in north central Sonora; Brown collected there in May and June 1905 for the late John E. Thayer. Col. Thayer gave some of the birds to Outram Bangs and they became a part of the Museum of Comparative Zoology collection when the museum ac- quired the Bangs Collection. Aimophila rufescens subvespera Griscom Aimophila rufescens subvespera Griscom, Bull. Mus. Comp. Zool., 75, 1934 , p. 418. Type. No. 164571, ad. 9 ; Guerrero: Chilpancingo; 2 March, 1932; W. W. Brown. Aimophila ruficeps simulans van Rossem Aimophila ruficeps simulans van Rossen, Bull. Mus. Comp. Zool., 77, 7, 29 December, 1934, p. 486. Type. No. 222783, ad.cT; Chihuahua: Mina Abundancia; 20 April 1888; M. Abbott Frazar. Passerculus sandwichensis oblitus Peters and Griscom Passerculus sandwichensis oblitus Peters and Griscom, Bull. Mus. Comp. Zool., 80, 13, 19 January, 1938, p. 454. Type. No. 172949 (formerly no. 23851 National Museum of Canada), ad. d" ; Manitoba: Churchill; 4 June, 1930; P. A. Taverner. Passerculus sandwichensis crassus Peters and Griscom Passerculus sandwichensis crassus Peters and Griscom, Bull. Mus. Comp. Zool., 80, 13, 19 January, 1938, p. 459. Type. No. 322033 (formerly no. 22033 J. E. Thayer Collection),