MID-PL.EISTOCENE FROM WESTERN NEBRASKA, INCLUDING A NEW SPECIES OF SHELDGOOSE

LESTER L. SHORT, JR.

American Museum of Natural History New York, New York 10024

I have studied a number of avian fossils from ontological work (see discussion by Cracraft western Nebraska, originally borrowed for 19683-4). study by Charles G. Sibley from the University of Nebraska State Museum. Among the fossils LIST OF SPECIES are 10 elements from the Rushville Quarries, cumxknsis? Canada Goose. The prox- 14 mi. S and 2 mi. W of Rushville, Sheridan imal third of a left ulna (U.N.S.M. 5768) County, Nebraska. The fossils represent six of a large goose closely approximates that of extant and one new fossil species, all but one the largest specimens of Canada Goose avail- of which are waterfowl. Their age is Yarmouth able to me. The papillae for attachment of the Interglacial (probably early Yarmouth) of the secondaries are farther apart than in Recent middle Pleistocene. Canada Geese, but this appears to reflect its The precise localities represented are Uni- large size ( slightly larger than an ulna of B. c. versity of Nebraska State Museum (U.N.S.M.) occidentalis, from the largest available speci- collecting locality Sh-3, NE W, NW % sec. 9, men of Canada Goose). Despite the close T. 29 N., R. 44 W., and collecting locality Sh-4, resemblance of the fossil to ulnas of Canada NW %, NE %, sec. 9, T. 29 N., R. 44 W. The Geese, I only tentatively assign the ulna to Rushville Quarries were described and their species because the ulna is not a particularly history and associated mammal fauna listed by diagnostic bone in the ( Woolfenden Schultz and Tanner (1957). Locality Sh-3 is 1961). Measurements of the ulna are: trans- illustrated in Schultz and Stout (1961). The verse distance from olecranon process to the fossils described herein are pale brown in edge of the external cotyla, 21.5 mm; and shaft color and are derived from Sappa sand in width across prominence for anterior articular Terrace-4 fill (Schultz and Tanner, op. cit.). ligament, 14.7 mm. Since Sangamon soil is exposed “high above Anubernicula Ross, new species of sheld- the quarry level” (Schultz and Stout 1961:X)), goose, described as follows: the fossils are definitely pre-Sangamon; the Sappa sand is pre-Illinoian. Fossils are partic- Anabernicula robusta, new species ularly abundant in a narrow band of fine, light Figures 1, 2 brown sand low in the Sappa formation, placed Type. Right humerus, complete except for as early Yarmouth in a section of the (Sh-3) part of the deltoid crest and the distal part of Rushville Quarry by Schultz and Stout ( 1961: the bicipital crest, U.N.S.M. 5769. fig. 36). According to these authors, the only Locality and Age. U.N.S.M. Collecting other fossiliferous portion of the Sappa forma- Locality Sh-3, Rushville Quarries, Sheridan tion is a white marl segment situated much County, Nebraska (see locality information higher than (hence later than) the above- above). Age-Yarmouth Interglacial (proba- mentioned fine band of sand. None of the bly early) of Pleistocene. fossils reported herein exhibits the light gray Diagnosis. Resembles humeri of Anaberni- and white color that would be expected of cub minuscula (Wetmore), A. gracilenta Ross, fossils from the marl portion of the Sappa and A. oregonend Howard, but more robust formation. All the fossils were collected be- (see below); anterior base of pneumatic fossa tween 1933 and 1938, mainly by C. B. Schultz, raised, forming a strong shelf, and causing the T. M. Stout, and their associates of the Uni- pneumatic fossa to be distinctly angular, versity of Nebraska State Museum. rather than circular in shape (viewed in- The osteological terminology employed temally) as in other species of Anabedculu; herein follows that of Howard (1929) and generally most similar to the humerus of the Ashley ( 1941). Stereophotography is em- Pliocene A. minusculu (Wetmore 1924)) but ployed because it is effective in osteological with a deeper ligamental furrow (nearer A. comparisons, and I believe it facilitates pale- oregonerzsis, Howard 1964), the shaft ridge

D471 The Condor, 7.2: 147-152, 1970 146 LESTER L. SHORT, JR.

- FIGURE 1. Stereophotographs of the right humerus of Neochen jubata (left set), A.M.N.H. no. 2545, and Anabemicukz robusta (right set), U.N.S.M. no. 5769, type specimen. Posterior view. Approximately natural size.

more external and less strongly marked, the Specimens examined. These include the shaft more strongly curved, and the head holotype humeri of Anaberniculu minuwula rotated more anconally; differs from A. gruci- (U.S.N.M. no. 10548) and of A. oregonen& Zen& (Howard in Delacour 1964) and A. ( A.M.N.H. no. 3548). oregonensis in its much more robust shaft, its Remarks. The new fossil humerus does not proportionally wider condylar end, its more fit within the closely similar A. oregone&-A. prominent attachment of anterior articular grucikntu complex (see Howard 1964:5, lo), ligament (which agrees with with of A. graci- nor is it in any way intermediate between the Zentuin facing more externally and less palmad humeri of these late Pleistocene species and than in A. oregonerwis),its somewhat higher the Pliocene A. minuscula. It appears to head, and its more externally placed shaft represent a rather stockier “pygmy goose” of ridge, resulting in a much greater depression a size similar to these species. between the head, the median crest and the Howard (1964:8) has properly placed shaft ridge. Anubernicula in the Tadomini, comparing the Measurements.Length, 98.1 mm; breadth of fossil specimens with Tadorna. Howard (in head from external tuberosity to the bicipital Delacour 1964:286) had stated previously, crest, 21.5 mm; breadth of distal end, 16.8 mm; “the placement of the genus (Anubernicula) breadth of shaft in center, 8.5 mm; depth of in this tribe (Tadomini) should be considered shaft in center, 7.4 mm. (Comparable mea- tentative pending comparisons with other surements obtained from the type humerus of genera of the tribe.” I have compared An- A. megonensis are, respectively, 98.1 mm, 20.0 aberniculu robusta and A. oregonen& humeri mm, 14.7 mm, 6.8 mm, and 6.0 mm.) with those of Cereopsis, Chloephuga,Neochen, MID-PLEISTOCENE BIRDS FROM WESTERN NEBRASKA 149

FIGURE 2. Stereophotographsof the right humerus of Neochen jubatu (top left set), A.M.N.H. no. 2545, and AnabertzicuZurobusta (top right set, bottom set), U.N.S.M. no. 5769, type specimen. Anterior view at top, distal end at bottom. Approximately natural size.

Alopochen, Tadornu (including Ca.sarca) and anconally so that the external tuberosity is Tuchyeres, as well as humeri of species of high and prominent; the anterior articular liga- other tribes of the Anatidae. This comparison mental facet raised; and the shaft tending clearly indicates that Anaberniculu belongs in to be relatively thin. (In this respect An- the Tadornini, and that it is closely allied to abernicula robusta tends toward stockier the Chloephuga-Alopochen-Neochen group of anatids like the Cairinini.) Woolfenden (1961: that tribe. Humeri of Anuberniculu are most 11) considers Neochen to be quite different similar to those of Neochen iubata (see figs. from Alopochen, Chloephuga, and other shel- 1 and 2). , primarily because the depression for Woolfenden ( 1961) has discussed the fea- the external head of the triceps extends to the tures of the humerus characterizing various external tuberosity in Neochen jubata, instead genera of Tadomini. Anuberniculu agrees with of being separated from it by a raised area. the Tadomini in the following features: However, this region in Neochen is not very capital shaft ridge fairly prominent and different from the narrow raised area found in directed toward the external tuberosity; various species of Chloephaga. In fact, deltoid crest extending quite far distally Chloephuga can be considered intermediate relative to the bicipital crest; the head rotated between Neochen and Alopochen in this 150 LESTER L. SHORT, JR. respect. I consider the humeri of Neochen to fossil. The greatest distance diagonally across be very similar in all respects to Chloephaga, the shaft of the fossil ulna from the internal and note that Johnsgard (1965) considers trochlea to the external trochlea is 10.5 mm; Neochen closely related to Chloephaga on be- the depth of the bone at the proximal base of havioral grounds. its external trochlea is 6.9 mm; and the greatest It might be useful to distinguish humeri of diagonal distance across the external trochlea Anabernicula from those of Chloephaga and is 10.1 mm. Neochen. Compared with humeri of Chloe- Spat& clypeata? Shoveller. The distal half phaga~ and Neochen, those of Anabernicula of a left tarsometatarsus (U.N.S.M. 5771) is have: shallower heads; a narrow raised area that of a medium-sized anatine . Its between the external tuberosity and the de- trochlea for digit II is short and extends pression for attachment of the external head distally beyond the level of the proximal of the triceps (A. oregonensis varies in this extent of the facet of the trochlea for digit III, regard from the condition of Neochen to that which feature separates (Woolfenden 1961: of Chloephaga); a lower shelf at the anterior 81) the non-diving from the diving . base of the pneumatic fossa (the shelf is well Comparison of the fossil with tarsometatarsi developed in A. robusta, which closely ap- of all North American species of Anus (in- proaches Neochen in this feature, and in the cluding Mareca) and SpatuZa disclosed that it correlated shape of its pneumatic fossa); less matched Spatula clypeata. I could find no robust and less pronounced trochleae; abrupt features of size, proportions, details of troch- contact of the bicipital crest with the shaft lear arrangement, or facets and articulations (crest tapering at an angle in Neochen) ; and by which the fossil differed from S. clypeata. the shaft distance between level of the distal In view of the great similarity among species base of the bicipital crest and level of the of Anas, it seems prudent to designate the distal base of the deltoid crest less (Neochen partial fossil tarsometatarsus only tentatively is intermediate in this respect, while Chloe- as Spat& clypeata. Some measurements are: phaga resembles Alopochen). Finally, the greatest width across trochleae, 7.5 mm; dis- humerus of Neochen has a distinct ridge at tance from proximal base of trochlea II to the the distal end of the olecranal fossa; this ridge distal end of trochlea III, 8.9 mm; and breadth is lacking or barely suggested in Anabernicula of shaft at base of trochlea II, 4.4 mm. and Chloephaga. Lophodytes cucullatus. Hooded Merganser. While I cannot comment on the status of The distal three-quarters of a right coracoid Brantadorna, a recently described (Howard (U.N.S.M. 5766) is identical in size, shape and 1963:8-g) genus assigned to the Tadomini, it details of conformation with coracoids of this differs from Anabernicula (including A. species. It matches Lophodytes in features robusta) in the features of its humerus. cited by Woolfenden ( 1961: 56-57)) particu- Howard stated that in all the characters by larly the shape of the glenoid facet, used to which elements of Anabernicula differ from separate Lophodytes from Mergus. Measure- Brantadorna, the latter “is more gooselike than ments are as follows: distance from procoracoid is Anabernicula.” to anterior end of bone, 10.7 mm; and maxi- Anas platyrhynchos? Mallard. The distal mum breadth of the bone through the glenoid half of a left ulna (U.N.S.M. 5767) compares facet, 7.0 mm. favorably with those available for large males Mebagris gallopavo? Turkey. The distal of this species. A slight difference from the 60 per cent of a left tarsometatarsus (U.N.S.M. Mallard is that the ridge of the external 5773) lacking trochlea III, and poorly pieced trochlea is very short for the large size of the together along the shaft and at both remaining shaft, which slopes more internally than in trochleae, appears to represent this species, most Mallards. The proportionally shorter but its poor condition renders this determina- ridge of the external trochlea suggests the ulna tion tentative. The specimen is also consider- of the Pintail (Anas acuta). However, the ably abraded. Its lack of a spur indicates that fossil ulna is too large and robust for Anas it represents a female, further complicating acuta, from which it differs also in having a the determination, since characters involving raised area at the proximal base of the internal the spur ( Brodkorb 1964b:225) are important trochlea (dorsally), as in all Mallards ex- in delimiting genera of turkeys. The fossil amined. Although it is likely to represent a tarsometatarsus resembles that of Meleagris Mallard, the relatively non-diagnostic char- (spp.) in the following characters discussed acter of anatid ulnas is reason for caution in by Brodkorb (lot. cit.) : lateral distal foramen definitely assigning a specific name to the high and inner trochlea narrow. Its facet for MID-PLEISTOCENE BIRDS FROM WESTERN NEBRASKA 151 the hind toe appears to be rather high, but trochlea II and distal end of trochlea III, 11.3, it is virtually indiscernible. Its inner distal 11.3, 11.9, and 10.3 mm; breadth of shaft at foramen is small, but well developed. Al- proximal base of trochlea II, 5.8, 5.9, 5.8, and though I have not compared the specimen 5.0 mm; total length, 59.9 mm; and depth of with tarsometatarsi of other fossil turkeys head through hypotarsus, 9.7 mm. (Agriocharis progenes, A. leopoldi, A. cras- sipes, Parapavo californicus, Meleagris alta, and DISCUSSION M. tridens are fossil turkeys represented by These fossils suggest for their habitat a marsh- tarsometatarsi, according to Brodkorb 1964a), edge ecological situation in a climate not careful comparison with numerous prehistoric greatly different from that of today. The oc- and modem tarsometatarsi of M. gallopavo currence of the Turkey suggests the presence places the fossil well within the range of varia- of trees, as does the occurrence in the Rush- tion in females of this species. Although the ville Quarry deposits of raccoons (Prucyon fossil tarsometatarsus is indistinguishable from sp. ), giant beavers ( Ca.storoides sp. ) and that of M. gallopavo, in view of its poor con- beavers (Castor sp. ) (Schultz and Tanner dition it seems best to assign it only tentatively 1957:71). Also found in these deposits and to this species. The maximum width across suggesting a wet environment is a muskrat its trochleae is 25.0 mm, and the shaft at about (Ondutra nebrascensis, Schultz and Tanner, its center measures 10.2 mm in width and 8.0 lot. cit.), Other mammals from this site are mm in depth. various horses, a camel (Camelops sp.), a Fulica americana American Coot. Three llama (Tanupolama sp.), several species of left tarsometatarsi, including a complete one, pronghoms, a vole, a prairie dog, a ground the distal one-third of another, and the distal squirrel, and a jack rabbit, all of which in- 40 per cent of the third (respectively U.N.S.M. dicate the occurrence of prairie habitat. Ex- nos. 5772, 5770 and 5765), are definitely as- cept for the sheldgoose, all avian species signable to this species. The distal one-quarter reported herein presently occur, or have re- of a smalIer right tarsometatarsus (U.N.S.M. cently occurred, in Nebraska (A.O.U. 1957). 5776) probably also represents this coot. All The extinction of sheldgeese ( Anubernicula) of these agree with Fulica and Gallinula, as in North America is peculiar, especially con- opposed to Rallus, in their flattened anterior sidering that their South American relatives shaft and well-spread trochleae (see Ligon apparently thrive today in both tropical 1965: 141). The three large tarsometatarsi ( Neochen jubata) and temperate (Chloe- closely match Recent tarsometatarsi of this phaga spp. ) areas. species, although their trochlea IV is slightly smaller in size. The small specimen is peculiar ACKNOWLEDGMENTS in having relatively smaller trochleae II and IV, but this may be attributable to excessive I am grateful to the authorities of the U. S. National Museum (U.S.N.M.) and the American Museum of abrasion. There is a possibility that the latter Natural History (A.M.N.H.), where these studies fossil may represent Gallinulu chloropus, the were accomplished. I thank C. G. Sibley for turning tarsometatarsi of which have smaller inner and the fossils over to me for study, and C. B. Schultz and outer trochleae than do those of F. americana. H. L. Gunderson of the University of Nebraska State Museum for the loan of the material and for much (Available comparative material included 10 helpful information. J. Cracraft discussed aspects of tarsometatarsi of modern F. americana and the study with me, to the benefit of this paper. H. four of modem Gallin& chloropus; critical Howard provided specimen data and helpful criticism data from five additional specimens of the of the manuscript. former and two of the latter were kindly pro- LITERATURE CITED vided by H. Howard.) However, the similarity of their tarsometatarsi, including overlap in AMERICAN ORNITHOLOGISTS ’ UNION. 1957.’ Check- measurements, makes difficult the determina- list of North American birds. Fifth ed. A.O.U., Baltimore. tion of the specimen. The incompleteness and ASHLEY, J. F. 1941. A study of the structure of the abrasion of the fossil tarsometatarsus in ques- humerus in the Corvidae. Condor 43:184-195. tion and the occurrence of three coot tarso- BRODKORB,P. 1964a. Catalogue of fossil birds, part metatarsi in the fossil material from this 2. Bull. Florida St. Mus. 8:195-335. locality make it prudent to assign it tenta- BRODKORB, P. 196413. Notes on fossil turkeys. tively to F. americana. Measurements of these Quart. J. Florida Acad. Sci. 27:223-229. fossils follow in the order presented above: CRACRAFT, J. lQ68. A review of the Bathornithidae ( Aves, Gruiformes), with remarks on the re- width across trochleae, 9.9, 9.9, 9.4, and 8.7 lationships of the suborder Cariamae. Amer. mm; distance between proximal base of Mus. Novitates, no. 2326. 152 LESTER L. SHORT, JR.

DELACOUR, J. 1964. Waterfowl of the World. Vol. Haile, Florida. Bull. Florida St. Mus. 10:127- 4. Heinman, New York. 158. HOWARD, H. 1929. The avifauna of the Emery- SCHULTZ, C. B., AND T. M. STOUT. 1961. Field ville shelhnound. Univ. California Publ. Zool. conference on the Tertiary and Pleistocene of 32:301-394. western Nebraska. Univ. Nebraska St; Mus., HOWARD, H. 1983. Fossil birds from the Anza- Spec. Publ. no, 2. Borrego Desert. Los Angeles Co. Mus. Contrib. SCHULTZ, C. B., AND L. G. TANNER. 1957. Medial Sci., no. 73. Pleistocene fossil vertebrate localities in Ne- HOWARD, H. 1984. A new species of the “Pigmy braska. Bull. Univ. Nebraska St. Mus. 4:59-81. Goose,” Anubernicukz, from the Oregon Pleisto- Wrrr~onz, A. 1924. Fossil birds from southeastern cene, with a discussion of the genus. Amer. Mus. Arizona. Proc. U. S. Natl. Mus. 64:1-18. Novitates, no. 2200. WOOLFENDEN, G. E. 1981. Postcranial osteology JOHNSCARD, P. A. 1965. Handbook of waterfowl of the waterfowl. Bull. Florida St. Mus. 6:1-129. behavior. Cornell Univ. Press, Ithaca, New York. LICON, J. D. 1965. A Pleistocene avifauna from Accepted for publication 28 February 1969.