9 November Í979 • Vol. 206 • No. 4419 SíWlafííW AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE

Ts, covín Reconstruction of the largest species in the family Plotopteridae, shown to scale with the outline of the largest living . These giant in- habited the North Pacific about 30 mil- lion years ago. They were flightless and had paddle-like wings similar to pen- guins, but belong to the unrelated order Peiecaniformes. See page 688. [B. Dal- zell, Rockville, Maryland] Fossil Counterparts of Giant from the North Pacific

Abstract. New of giant, flightless penguinlike have been found in late Oligocène and early rocks in Japan and in the state of . These birds belong to the order Pelecaniformes. in the extinct family Plotopteridae, pre- viously known by a single fragment of bone from California. Hindlimb and pelvic morphology is most similar to that of Recent anhingas. but the wing is paddlelike and remarkably convergent toward penguins and flightless auks. Both the Plotopteridae and the giant penguins became extinct by the middle Miocene, possibly because of competition from seals and porpoises.

Fossils have recently revealed an un- , anhingas, and allies). They known avian family which includes some are known only from mid-Tertiary rocks of the largest swimming birds yet discov- bordering the North Pacific. ered. These marine birds were flightless, Knowledge of these birds began with a wing-propelled divers resembling pen- single humeral end of a coracoid de- guins (Sphenisciformes) in their locomo- scribed from an early Miocene deposit in tory adaptations but belonging to the un- southern California as a new and related order Pelecaniformes (pelicans. species, Plotopterum joaquinensis (/). With exceptional insight, Howard (¡) di- agnosed this as a new family of Pelecani- formes, the Plotopteridae, having affini- ties with anhingas and cormorants (An- hingidae, Phalacrocoracidae) but show- ing adaptations for wing-propelled diving similar to those of auks (Charadrii- formes, Alcidae) and penguins. How- ever, because of its very fragmentary na- ture, Plotopterum drew little attention. Newly discovered fossils fully confirm Howard's original conclusions. All are late Oligocène or early Miocene {2,3) and consist of single elements or associ- ated partial skeletons from six localities in Japan and one partial skeleton from the state of Washington. Of the major skeletal elements, only the end of the bill remains unknown. Coracoids associated with the speci- men from Washington and with two from Japan clearly show that the new fossils are referable to the Plotopteridae. Counting Plotopterum, at least three Fig. 1. Dorsal view of right wing skeleton. (A) genera can be recognized (J). Dif- Anhinga (Pelecaniformes); (B) great auk (Cha- ferences in size indicate a minimum of radriiformes); (C) plotopterid (Pelecani- four species from Japan, and the Wash- formes; largest Japanese species); and (D) ington specimen could represent a fifth. penguin (Sphenisciformes). The similarities Considerable taxonomic diversity within between the three wings on the right are due to convergence. The plotopterid (C) evolved the family is indicated. from an ancestor with a wing like that of the Plotopterum, by far the smallest anhinga (A). Drawn to scale. known member of the family, was the

Fig. 2. Anterior view of right tarsometatarsus. (A) Anhinga (Pelecaniformes); (B) ploto- pterid (Pelecaniformes; largest Japanese species): and (C) penguin (Sphenisciformes). Not to scale.

c 0036-8075/79/1109-0688ÎOO.ÎO/0 Copyright © 1979 AAAS SCIENCE, VOL. 206, 9 NOVEMBER 1979 size of the modern Phaia- of birds. The penguins are so modified to multaneous disappearance adds consid- crocorax peniciUatiis (I). Other species this end that their ancestry is no longer erable support to Simpson's hypothesis. were only slightly smaller than the larg- readily discernible. Nevertheless, de- In addition, the Plotopteridae provide est living penguins, and two species far spite earlier opinions to the contrary, one of the more impressive instances of exceeded existing penguins in size. In they are now generally regarded as hav- in the fossil record. life, the largest Japanese plotopterid (see ing descended from volant ancestors (6). STORRS L. OLSON cover) probably measured more than 2 m The convergent adaptations of the Plo- Department of Vertebrate Zoology, from bill tip to tail tip and may have been topteridae offer significant proof that al- National Museum of Natural History, larger (4) than any of the giant fossil pen- most any group of water birds could Smithsonian Institution, guins from the Tertiary of the Southern evolve an essentially penguinlike mor- Washington, D.C. 20560 Hemisphere (5). phology in becoming specialized for YOSHIKAZU HASEGAWA The most striking feature of the Plo- wing-propelled diving. National Science Museum, topteridae is the wing, which had lost all Storer (6) remarked on the absence in Tokyo ¡60, Japan resemblance to that in other Pelecani- the Northern Hemisphere of wing-pro- formes and was modified as a rigid pad- pelled diving birds of the size of the References and Notes dle, unsuitable for flight. The wing bones larger Recent penguins or their giant fos- L H. Howard, Condor 71. 68 (1969). 2. The Washington specimen is from the Pysht show numerous parallels with those of sil relatives. We now know, however, Formation. Twin River Group. Japanese speci- penguins and the flightless Alcidae (Fig. that the Plotopteridae occupied such mens are from the Asagai, Ashiya, Hatazu, and Nishisonogt formations [see Y. Hasegawa, 1). The head of the humérus is heavy, niches in the North Pacific for an unde- Proc. Is! In!. Congr. Far, Neogene Siraiigr. nearly spherical, and penguinlike, termined period prior to the end of the (1977(, p. 340; Kagalcu Asahi Î. 71 (1978); BuU. Kita Kyushu Mus. Mal. Hisi. 1, 41 (1979)]. whereas the distal end of the humérus early Miocene. Plotopterids are not 3. S. Olson. Nal, Hist. Mus. Los Angeles Cty. and the elongated first metacarpal more Conirib. Sei.. in press. known from the more intensively studied 4. The length of a recently collected synsacrum of closely resemble those of flightless auks. younger deposits around the North Pa- Anthropornis nordenskjoeidii. the largest or nejit to largest of the giant fossil penguins, is 240 A row of distinct pits on the dorsal sur- cific, and it seems likely that they be- mm. whereas that in the targes! Japanese plo- face of the ulna is unique; in most birds came extinct at the same time as the gi- topterid is approximately 330 mm. 5. G, Simpson, in The Biology of Penguins, B. the secondaries attach to raised papillae. ant penguins in the Southern Hemi- Stonehouse, Ed. (Macmillan. Lxindon. 1974), The articulating surfaces of the wrist sphere. Simpson (5) has suggested that pp. 30-32. 6. R. Storer, Proc. 12th Inl. Ornitkol. Congr. bones indicate that the hand was capable niches for pelagic endotherms the size of (1960), p, 694. of very little flexion (J). giant penguins may have been pre- 7. For collecting, preparing, or providing informa- tion on specimens we thank U. Akagi, H. Endo, As in penguins, the scapula was great- empted by seals and porpoises, which D. Emlong, S. Isotani, K. lso^akl, Kiu Kyusyu Natural History Museum. H. Kodama, K. Mat- ly expanded to provide increased attach- underwent their greatest diversification suo. J. Mori, S. Ota, H. Otsuka, C, Ray. K. ment for the muscles used to raise the during the Miocene. Because plotopte- Seki. G. Sullivan, Y. Suyana, T. Suzuki, and W. Zinsmeister. Illustrations are by B. DaJzell. J. wing against water, a much denser medi- rids occupied niches similar to those of Farrand, C. Ray, and D. Steadman commented um than air. Also as in penguins, the giant penguins but are unrelated and oc- on the manuscript. limb bones were extremely dense and curred in a difl^erent hemisphere, their si- 13 April 1979; revised 14 May 1979 heavy, serving to reduce buoyancy. Despite the great modification of the wing and scapula, the remainder of the skeleton in the Plotopteridae shows close affinity with the Pelecaniformes, The carina of the sternum articulates sol- idly with the furcula, there is a very large acromion process of the scapula, and the skull has a deep, transverse nasofrontal hinge and lacks supraorbitai impressions for salt glands. These characteristic fea- tures of the Pelecaniformes are not found in penguins or auks. The tarsometa- tarsus, although quite robust, is most like that of anhingas and has little resem- blance to that of penguins, in which, for example, the metatarsals are incom- pletely fused (Fig. 2). In much of the skeleton of the Plo- topteridae. the greatest similarity is to anhingas, which otherwise differ consid- erably in being freshwater, foot-pro- pelled diving birds with the head and neck highly modified as a spearing de- vice. Such specializations of the head and neck are absent in the Plotopteridae, however, and the skull is more like that in the Sulidae ( and ). Adaptations for the use of the wings rather than the feet in underwater pro- pulsion have evolved in several groups

SCIENCE, VOL. 206, 9 NOVEMBER 1979 689