248 SHORT COMMUNICATIONS kunthii blossoms by Bombus queens occurred and depend on many factors. That hummingbirds and the workers were unable to secure nectar while positioned ancestor of P. kunthii co-existed may be assumed; within the floral tube, probably as much as lo-20 otherwise its adaptation to hummingbird pollination per cent more nectar was available to Bombus p&her would make little sense. Thus it is possible that P. and Bombus trinominatus populations during this kunthii could have undergone much of its development period due to the feeding activity of Diglossa. under selective pressure from hummingbirds; still, it is clear that Diglossa baritula has co-existed with DISCUSSION AND CONCLUSIONS hummingbirds throughout New World montane hab- Grant and Grant (1968) have proposed an explanation itats for some time and therefore an earlier and more for the reciprocal evolution of hummingbirds and the important role in the evolution of P. kunthii would plants upon which they feed. According to this inter- not be unexpected. This is not to suggest that exploita- pretation most hummingbird-pollinated flowers, espe- tion late in the evolutionary development of P. kunthii cially temperate species, have evolved from bee flowers would be insignificant. Even at present, given the (Grant 1961; Grant and Grant 1965). The process potential counter-selection pressures on P. kunthii involves an incipient stage during which a primitive from bees, the presence of Dglossa perforations un- hummingbird or progenitor already “preadapted” to doubtedly precludes a certain amount of bee pollina- feed on a particular bee flower (in the sense of tion which would probably otherwise occur, helping securing insects within the corolla, or nectar, or both), to maintain the selection pressures on P. kunthii which causes at first occasional but then increasingly frequent favor continued floral specialization for hummingbird pollinations. This stage is followed by a resultant pollination. eventual elongation (and presumably a decrease in ACKNOWLEDGMENTS diameter) of the floral tube and thus an increasing exclusiveness for bird pollination. The following dis- Financial assistance for this study was provided by cussion assumes that the evolutionary history of P. National Science Foundation Grant GY8-4386. The kunthii follows the model, i.e., that this Mexican high- Department of Zoology, Entomology Section, Iowa land species has evolved from a relatively unspecial- State University, kindly made the identifications of ized bee-pollinated form to its present condition as a Bombus. We are esneciallv indebted to Robert Cru- A , hummingbird-pollinated species with a high degree of den, University of Iowa, for his many comments and exclusiveness. suggestions during preparation of the manuscript. More specifically, the Diglossa-Bombus exploitation could aid in the selection of P. kunthii as a bird LITERATURE CITED flower, as indicated in the following chronological GRANT, V. 1961. The diversity of pollination sys- schema. tems in the phlox family. Recent Advan. Bot. 1. Initial increase in nectar production is favored ( Toronto) 1: 55-60. by hummingbird visitation concomitant with begin- GRANT, V., AND K. A. GRANT. 1965. Flower pol- nings of corolla tube elongation due to increasing lination in the phlox family. Columbia Univ. pollination by hummingbirds compared with bees (i.e., Press, New York and London. birds visit those plants with the greatest nectar supply). GIIANT, V., AND K. A. GRANT. 19868. Humming- 2. Nectar production becomes sufficient to en- birds and their flowers. Columbia Univ. Press, courage exploitation by Diglossa. New York and London. 3. Perforations attract Bombus spp., diminishing MACIOR, L. 1966. Foraging behavior of Bombus bee pollination and thus decreasing or removing bees (Hymenoptera: Apidae) in relation to Aquilegia as a selective force. pollination. Amer. J. Bot. 53:302-309. 4. More rapid evolution of tube elongation and SKUTCH, A. 1954. Life histories of Central Amer- nectar production occurs, caused by a self-reinforcing ican birds. Pacific Coast Avifauna, No. 31. feedback system in which greater nectar production SPRENGEL, C. K. 1793. Das entdeckte Geheimnis encourages an increasing exploitation by Diglossa, der Natur im Barr und in der Befruchtung der which further discourages bee pollination relative to Blumen. In Ostwalds’ Klassiker der exakten Wis- bird pollination, enhancing further increase in exclu- senschaften, Nr. 48. 1894. Engelmann, Leipzig. siveness and nectar production. WOLF, L. L. 1969. Female territoriality in a trop- ical hummingbird. Auk 86:490-504. The stage in the evolution of P. kunthii at which Diglossa exploitation began is unknown and would Accepted for publication 11 November 1970. AN EARLY PLEISTOCENE Gigantocamelus fricki Barbour and Schultz (holotype). EAGLE FROM NEBRASKA An account of the stratigraphy and fauna of the Broadwater and Lisco localities is given by Schultz and Stout (1948). LARRY D. MARTIN The eagle is represented by the distal end of a University of Nebraska State Museum Lincoln, Nebraska 68508 tarsometatarsus and most of the shaft. The proximal end of a femur and the badly crushed distal end of a tibiotarsus was found associated with the tarsometa- In the collections of the University of Nebraska State tarsus and may belong to the same bird. The tarso- Museum (U.N.S.M.) are a number of fossil birds metatarsus is too large and massive to be satisfactorily from the Early Pleistocene Broadwater and Lisco compared with any of the North American Buteoninae, Local Faunas. Included with this collection is a new which also have the tarsometatarsus tapering more species of eagle from U.N.S.M. Coll. Lot. Gd-12. abruptly distally. It resembles the tarsometatarsus of Other animals from this locality are MegaZonyx, Aquila chrysaetos very closely, but differs from that Geomys, Procastoroides, Canis, Borophagus, Ischy- form in being more elongate and in having the papilla rosmilus, Mammut ( Pliomastodon), Plesippus, and for the tibialis anticus more proximally situated. It SHORT COMMUNICATIONS 249 FIGURE 1. Tarsometatarsus of Spizaelus tanneri, U.N.S .M. 20038, anterior, distal, and posterior views of type specimen, X 1. does resemble Spizuetus grinnelli (Miller) and S. When viewed anteriorly the trochlea for digit 2 ap- willetti Howard in these and other characters, and is pears large and extends further proximally than in the referred to Spizaetus. Haliaeetus leucocephdus has extinct species or in most specimens of Aquila. How- a much shorter, heavier tarsometatarsus, with the ever, when viewed posteriorly the trochlea for digit 2 papilla for the tibialis anticus more distally placed and appears weak as in other species of Spizaetus. the trochlae for digit 2 extending further distally than Measurements. Dimensions of the holotype of S. do the other trochleae. tanneri U.N.S.M. 20038: width of distal end, 24.3 Comparisons were made with tarsometatarsi of mm; width of shaft at the proximal end of the facet Haliaeetus leucocephulus (5), Buteo regdis (2), B. for metatarsal I, 13 mm; distance from the distal end ( Geranoaetus) melanoleucus (2)) Aquila chrysaetos of the trochlea for digit 3 to the middle of tubercle ( 15 ), and Spizaetus tyrannus ( 1) . In addition, femora for tibialis anticus, 88.5 mm. (This measurement is were examined of Aquila chrysaetos (5), Spizaetus approximate because of the crushed condition of the tyrannus (l), and S. nipalensis ( 1). Published descrip- proximal end of the bone.) tions and illustrations were relied upon for the charac- Etymology. The species is named in honor of Lloyd ters of Spizaetus grinnelli (Howard 1932) and S. Tanner who supervised many of the excavations at willetti ( Howard 1935). the Broadwater and Lisco Localities and who has contributed much to our understanding of the Pleisto- Spizaetus tanneri, new species cene of the Central Great Plains. Figure 1 DISCUSSION Holotype. Distal end and most of the shaft of the left tarsometatarsus, U.N.S.M. 20038. Spizaetus tunneri differs from Aquila in having the Type locality and horizon. U.N.S.M. Coll. Lot. tarsometatarsus more elongate and the papilla for the Gd-12, near center of E%SE1/9 sec. 13, T. 18 N, R. tibialis anticus more proximally placed. However, the 46 W, 3 mi. E, 2 mi. N of Lisco, Garden County, shaft is shorter and stouter than it is in Morphnus Nebraska. From the Broadwater Formation, Early_ woodwardi Miller. The shaft above the trochleae does Pleistocene. not narrow as much as in Aquila. Spizaetus tyrannus Referred material. Proximal end of femur U.N.S.M. also has the shaft of the tarsometatarsus less con- 20039 and the distal end of a tibiotarsus U.N.S.M. stricted above the trochleae than it is in Aquila. It 2#0040. This material was found associated with the resembles S. tunneri and differs from Aquila in having holotype and is probably part of the same individual. the medial margin of the shaft relatively straight, in Diagnosis. Tarsometatarsus slightly smaller than in having the shaft less indented at the facet for meta- Spizaetus willetti Howard from which it also differs tarsal I, in having the distal foramen smaller, and in in having the facet for metatarsal I not so large or so having the extensor groove leading into the distal lateral and the trochleae much more arched. It is foramen relatively deeper. The trochleae of S. tanneri larger and much more massive than S. g&me& (Mil- are arched more as in Aquila than as in Spizaetus ler) and has a wider shaft of the tarsometatarsus. willetti or Haliaeetus leucocephalus. The external 250 SHORT COMMUNICATIONS flange on the trochlea for digit 4 is slender as in Certainly it is useful to unite them in a single sub- Aquila and Spizaetus, and is not short and stubby as family, Aquilinae, as was suggested by Howard (1932). in Morphnus woodwardi (Howard, Condor 37:208, I am grateful to J. Cracraft for the loan of Recent 1935). The facet for metatarsal I is larger and higher skeletal material from the American Museum of than it is in Aquilu of similar size.