SUMMARY The fossil species Rhegminornis calobates, based on a fragmentary tarsometatarsus from the Lower Miocene deposits at Thomas Farm in central Florida, has previously been regarded as constituting either a distinct family within the Charadriiformcs or a subfamily of the Jacanidae. Re-examination of the type, and two additional specimens of the same element here assigned to the species, shows that Rhegminornis is actually a galliform referable to the family Meleagrididae. Rhegminornis thus becomes the oldest known turkey and is much smaller than modern members of the family. We suggest that the family Meleagrididae has had a fairly long and diverse history in North America. The allocation to the Cracidae of Boreortalis laesslei, another galliform bird from the Thomas Farm deposits, is confirmed. Additional cracid material from the same locality, possibly indicating the presence of more than one species is briefly noted.

The only fossil form referred to the Jacanidae is the Miocene Rhegminorns calobates Wetmore (1943), described from a distal end of a right tarsometa- tarsus. An interest in that family induced us to restudy this specimen, with the result that it proved to be neither a jacana nor a Charadriiform, but a galliform in the family Meleagrididae. This prompted us to examine other galliform material from the same locality--the Lower Miocene deposits at Thomas Farm, 8 miles north of Bell, Gilchrist County, Florida. DESCRIPTION AND COMPARISONS Rhegminorm's calobates was assigned to the order Charadriiformes, super- family Charadrioidea, by Wetmore (op. cit.). It was said to differ from the other groups in that superfamily in the relative positions of the trochleae and in the large scar for the hallux, indicating a much better-developed first toe. Wetmore created the family Rhegminornithidae for the species. He stated (op. cit.:6) that "the fossil shows a stage slightly intermediate between that found in the Scolopacidae and Charadriidae and the Jacanidae. The evident size of the hallux also is reminiscent of the latter family .... "This statement led Brodkorb (1967) to consider the species as forming a sub- family (Rhegminornithinae) of the Jacanidae. The type element of Rhegminornis (MCZ 2331) actually shows no re- semblance to the specialized tarsometatarsus of the Jacanidae, differing in the greater width of the shaft, in having the inner and outer trochleae much more elevated, the inner trochlea not as enlarged and bulbous, and in lacking the distinct spur on the dorso-lateral corner of the posterior face of the inner trochlea. The scar for the hallux is not as deep; the distal foramen is not located as far distally, is oval rather than circular, and is much smaller, with a much shallower outer extensor groove than in the Jacanidae. The measure- ments of the type are: width through trochleae 9.5 mm; depth of middle trochlea 4.6; least depth of shaft 2.9; least width of shaft 3.9; overall length of specimen 29.6. In the relative positions of the trochleae and large scar for the hallux, characters used by Wetmore to establish the family Rhegminornithidae, Rhegminornis agrees with typical Galliformes. In the following particulars as well, the type of Rhegminornis differs from the shorebirds and agrees with most of the Galliformes. Inner trochlea.--In posterior view (Fig. lb), the external side is flat and linear, not rounded nor extending laterally past the line of the internal side of the middle trochlea. 114 Olson and Farrand MIOCENE TURKEY 115 D FIc. 1. a, b, c. Stereophotographs of the tarsometatarsus of Rhegminornis calobates (holotype, MCZ 2331), about twice natural size (a. anterior view--inner intertrochlear foramen barely visible as a whitish spot; b. posterior view; c. distal view). d, e. Tar- sometatarsus of a small female Meleagrls gallopavo, natural size (d. anterior view; e. posterior view). f. Holotype of Rhegminornis calobates, natural size. June 1974 116 THE WILSON BULLETIN Vol. a6, No. 2 In internal view (Fig. 2a), the trochlea forms a distinct bulge distally, with the wing sharply set off from the articulating surface by a distinct pit; the roughened articular surface is much larger and extends farther proximally. In distal view (Fig. lc), the wing projects more medially rather than directly posteriorly and is more distinctly set off from the rest of the trochlea. As noted by Wetmore, the inner trochlea is not so strongly rotated towards the median axis of the shaft and is thus more nearly in the plane of the other trochleae. Middle trochlea.--In posterior view (Fig. lb), the articular surface is more distinctly set off from the shaft of the trochlea, especially on the internal side, and extends farther proximally on the external side. In internal view (Fig. 2a), the internal side is more excavated, with a more distinct rim. In external view (Fig. 3b), the external face bears a squared depression, unlike the deeper more circular pit of the shorebirds. External trochlea.--In posterior view (Fig. lb), the shape of the articular surface is quite different from that of the Charadrioidea, with the external side extending much farther proximally. In external view (Fig. 2b), there is a distinct groove on the postero- lateral portion of the trochlea to accommodate a ligament attaching the outer toe to the tarsus. This groove is a characteristic feature of the Galliformes but is absent in the Charadrioidea. These comparisons suffice to remove Rhegminornis from the Charadrii- formes and place it with the Galliformes. The question of its proper alloca- tion within the latter order may now be explored. Rhegminornis differs from the Cracidae and Megapodiidae in the following respects: inner trochlea much more elevated, middle trochlea rotated medially, external trochlea more elevated, distal foramen much larger, and the scar for the hallux with a single facet (as opposed to a distinct, polished, circular proximal facet and an elongated, less distinct, distal facet). From the Tetraonidae it differs in many characters, including the shape and positions of the trochleae, the relatively broader shaft with less expanded distal end, larger distal foramen, and larger scar for the hallux. Rhegminornis differs from the Phasianidae (sensu stricto) and agrees with the Meleagrididae in the more elevated internal trochlea, larger and more elongate distal foramen, and the presence of a distinct, although small, inner intertrochlear foramen. Howard (1927:24) noted that an inner intertrochlear foramen was present in more than half the specimens of Meleagris that she examined and that such a foramen, or traces of it, also occurred in Parapavo and 4griocharis but not in Pavo. In Rhegminornis the inner foramen is present both anteriorly (barely visible in Fig. la) and posteriorly (visible under magnification) in exactly the same position as observed in specimens of Meleagris (Fig. ld and e). We may add here that we did not find an inner intertrochlear foramen in the Charadrioidea. In the tarsometatarsi of Rhegminornis and the Meleagrididae, in side view, the middle trochlea curves anteriorly past the line of the shaft whereas in the other Galliformes, including the Phasianidae and the Numididae, the middle trochlea is more nearly in line with the shaft. Olson and Farrand MIOCENE TURKEY 117 C D F;. 2. Stereophotographs of tarsometatarsi of Rhegminornis calobates, about twice natural size. a. Holotype (MCZ 2331), internal view. b. Same, external view. c. Re- ferred prnximal end (PB 1776), anterior view. d. Same, posterior view. DISCUSSION The proper familial allocation of Rhegminornis appears to be with the Meleagrididae. It differs from all modern and fossil turkeys in its much smaller size (Fig. 1) and more elevated external trochlea, but is otherwise June 1974 118 THE WILSON BULLETIN Vol. 86, No. 2 very similar to modern forms. Rhegminornis agrees with Meleagris and differs from Agriocharis (as well as other Galliformes) in the more medially rotated middle trochlea. It more closely resembles Agriocharis, however, in having the trochleae less divergent and the distal foramen in posterior view smaller than in Meleagris (see Fig. 1). At least two other bones from the Thomas Farm deposits are referable to Rhegminornis. One of these (PB 8447) is another distal end of a right tarsometatarsus, much abraded and lacking the external trochlea; it is virtually identical to the type except that the scar for the hallux is more distinct. The other specimen (PB 8448) is an imperfect proximal end of a right tarsometatarsus (Fig. 2c and d) lacking much of the hypotarsus and part of the external cotyla. This differs from the tarsometatarsus of the Cracidae in the more excavated anterior shaft, the more medial position of the tubercle for tibialis anticus, the much more medially expanded internal cotyla, and particularly in not having the shaft greatly thinned and bladelike just distal to the internal cotyla as seen in the Cracidae. In all of these respects the specimen agrees very closely with the Meleagridi- dae. From Meleagris and ztgriocharis it differs principally in having the shaft somewhat less deeply excavated anteriorly and posteriorly, and in having the tubercle for the tibialis anticus narrower and more dearly defined. The thin, ossified intertendinal septurn extending down the shaft from the hy- potarsus in modern turkeys and certain other Galliformes is absent in Rhegminornis, but this variable feature is probably of little taxonomic sig- nificance. Likewise, there is no evidence of a spur in Rhegminornis, although this may only mean that the specimens were from females. From the same deposits at Thomas Farm, a new genus and species of cracid, Boreortalis laesslei Brodkorb (1954), was described. The type (PB 743) is a distal end of a right tibiotarsus lacking the internal condyle and the pos- terior portion of the external condyle. The fragmentary condition of the type and the fact that Brodkorb noted several differences between it and modern cracids suggested the possibility that Boreortalis might actually be referable to Rhegminornis. This, however, proved not to be the case. In the type of Boreortalis the remains of the internal condyle show that it was better de- veloped proximally than in turkeys, with a projection extending over the opening distal to the tendinal bridge--a decidedly cracid feature. Also, the tubercle on the tendinal bridge is more medially located in Boreortalis and most cracids than it is in the turkeys. Furthermore, there is additional unquestionable cracid material in the Thomas Farm deposits. A distal portion of a left tarsometatarsus (UF 2905) shows all the distinctive features of the Cracidae and indicates a species slightly smaller than Penelope superciliaris, as does a proximal end of a right Ol ...... d MIOCENE TURKEY 119 Farrand tarsometatarsus (PB 8450). A distal end of a left femur (PB 1776) is from a cracid very slightly smaller than Ortalis vetula. The type of Boreortalis laesslei was described as leing between Ortalis vetula and Peneloiina nigra in size (and is somewhat larger than Rhegminornis). It seems possible that more than one species of cracid is represented in the Thomas Farm deposits. In any event, Boreortal.:s is not a turkey and therefore not a synonym of Rhegminornis. Cracraft (1971) has assigned two distal ends of humeri from Thomas Farm to Boreortalis. In the better preserved of these specimens (MCZ 7068), the deeper olecranal fossa, the apparently broader and less curved shaft, much less distally projecting entepicondyle, and shallower brachial depression are more similar to Meleagris than to the Cracidae. On the other hand, it differs from modern turkeys in having no indentation letween the internal condyle and the entepicondyle and in having the brachial depression extending farther distally. These fragments, although slightly larger than might be expected for Rhegminornis, are possibly not properly assigned to Boreortalis. At any rate, we doubt whether any certain determination can be made of them until better material is available. Up to now, the oldest known turkey was Proagriocharis kimballensis Martin and Tate (1970), from the Upper Pliocene (Kimball formation) of Nebraska. The allocation of Rhegminornis to the Meleagrididae extends the fossil record of the family back much farther--into the Lower Miocene. The family otherwise consists of large forms in the genera Meleagris, etgriocharis, and Parapavo, that are represented by several Pleistocene, as well as two modern, species. Proagriocharis was smaller than any of the Pleistocene turkeys, and Rhegminornis was much smaller yet. The presence of a small but unmistakable turkey as early as the Lower Miocene suggests that the Meleagrididae have had a long history in North America and may well have undergone considerable radiation during the Tertiary. In view of this, it would perhaps be best to regard warily a number of the fossil taxa described from the North American Tertiary and currently assigned either to the Cracidae or the Tetraonidae. Some of these may well prove to be turkeys. As determined from study of the various vertebrate fossils from Thomas Farm, the environment in the area at the time Rhegminornis bones were de- posited there was one of a "river flowing through a dry grass-covered plain" (Brodkorb 1954:182). It is not difficult to envision a flock of diminutive turkeys, clucking and scratching its way along a sparsely wooded Florida riverbank. As the little flock moves on out of sight in the brush at the edge of the plain, a bantam-sized cock occasionally displays in Lilliputian splendor, strutting, gobbling and fanning its tail in the Miocene sunshine. June 1974 120 THE WILSON BULLETIN Vo. a6, No. 2 ACKNOWLEDGMENTS We are very much indebted to Raymond A. Paynter, Jr., Museum of Comparative Zoology (MCZ), for allowing us to study the type of Rhegminornis, and to Pierce Brodkorb for lending specimens from his own collections (PB) and those of the University of Florida (UF). In this study, as in many others, we have continued to enjoy the counsel and indulgence of our esteemed colleague, Alexander Wetmore. The photographs are by Victor E. Krantz. LITERATURE CITED BROOKORB, P. 1954. A chachalaca from the Miocene of Florida. Wilson BulL, 66: 180-183. BBovIzoaa, P. 1967. Catalogue of fossil birds: Part 3 (Ralliformes, Ichthyornithiformes, Charadriiformes). Bull. Florida State Mus., Biol. Sci., 11:99-220. CRACRAFT, J. 1971. The humerus of the early Miocene cracid, Boreortalis laesslei Brodkorb. Wilson Bull., 83:20201. HOWARD, H. 1927. A review of the fossil bird, Parapavo californicus (Miller) from the Pleistocene asphalt beds of Rancho La Brea. Univ. Calif. Publ. Bull. Dept. Geol. Sci., 17:1-62. MARTIN, L. D. AND J. TATE, JR. 1970. A new turkey from the Pliocene of Nebraska. Wilson Bull., 82:214-218. WETMOaE, A. 1943. Fossil birds from the Tertiary deposits of Florida. Proc. New England Zool. Club, 22:59-68. NATIONAL MUSEUM OF NATURAL HISTORY, SMITHSONIAN INSTITUTION, WASH- INGTON, D.C. 20560 AND DEPARTMENT OF ORNITHOLOGY, AMERICAN MUSEUM OF NATURAL HISTORY, NEW YORK, NEW YORK 10024. ACCEPTED 21 JANUARY 1974.