PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. 50, No. 4, pp. 1 15-138, 9 figs., 9 tables. December 9, 1997 A NEW LATE MIOCENE BALAENOPTERID WHALE (CETACEA: MYSTICETI), PARABALAl^(M^iUgiUJ^M/r^l\SIS, (NEW GENUS AND SPECIES) FROMf tffi^AW^fc^tfr ' MUDSTONE, POINT REYES PENINSULA, CALIFORNIA DEC 1 8 1997 By Carl V. Zeigler Woods Hole, MA 02543 Cetacean fossil technician, College of Marin, Kentfield, California 94904; deceased, 19DT.~~ Gordon L. Chan Professor of Biology, College of Marin, Kentfield, California 94904; deceased, 1996. and Lawrence G. Barnes Section of Vertebrate Paleontology, Natural Histoiy Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, California 90007 A fossil balaenopterid whale, Parabalaenoptera baulinensis, new genus and species, is named on the basis of a partial skeleton that was collected from the Late Miocene Santa Cruz Mudstone at Bolinas Point, Point Reyes Peninsula, California. Diatoms from the Santa Cruz Mudstone indicate an age of 6.8 to 6.0 Ma. The skeleton includes the cranium, dentaries, cervical and dorsal vertebrae, pectoral limb bones, and ribs, and the animal is estimated to have been 10 m in length when alive. The skull of Parabalaenoptera baulinensis has pronounced interdigitation of the rostral and cranial elements and an abrupt slope from the intertemporal region to the supraorbital process, indicating that it is a balaenopterid. However, the elongate and very narrow nasals, narrow intertemporal region, long and posteriorly tapering ascending processes of the maxillae, conspicuously elevated or swollen anterior portions of the parietals, short postglenoid processes, and long and sloping coronoid crest of the dentary distinguish P. baulinensis from fossil and Recent members of the subfamilies Megapterinae and Balaenopterinae. Its unique characters indicate its separation at the subfamily level, and we assign it to the new subfamily Parabalaenopterinae. Received September 11, 1995. Accepted April 15, 1996. Marine vertebrate remains occur within the sirenian, and Domning (1978) reported a du- Santa Cruz Mudstone exposed at the southern tip gongid sea cow from these exposures. From far- ofthe Point Reyes Peninsula, California (Fig. 1), ther north on the Point Reyes Peninsula, in the where rapid erosional retreat of the coastal cliffs same formation, Barnes (1976) reported a sperm frequently uncovers fossils. Galloway (1977) whale (cf. Scaldicetus du Bus, 1867); a small reported fragments of whale bone and an uniden- rorqual (cf. Burtinopsis Van Beneden, 1872); tified skull, possibly that of a cetacean or Wilson (1937) reported a dolphin; and Repen- [115] 116 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 ning and Tedford (1977) reported an imago- tariine pinniped (pseudo-sea lion). The Bolinas Point area has yielded two major discoveries, relatively near each other, only about 375 m apart. The dugongid sea cow re- ported by Domning (1978) was found in 1972. The second specimen, discovered 12 August 1976 by the senior author, began with the seren- dipitous discovery of a vertebra that had eroded from the cliff about 10 m above the beach. Fur- ther excavation along the cliff face yielded the partial skeleton of the balaenopterid which is the subject of this paper. We describe this fossil balaenopterid herein and name it as a new genus and species in a new subfamily of the family Balaenopteridae. MATERIALS AND METHODS The location of the holotype skeleton in the sea cliff made its excavation difficult. Collection was dangerous because of the steepness of the cliff face (10 m above the beach), frequent landslides due to the crumbly nature of the mudstone, and strong surf action against the cliff face during high tides. To solve the problem of recovery, we decided to remove the fossil bone as pieces were sequentially (and naturally) exposed in the rap- idly retreating cliff face. During preliminary excavation we found that the vertebral column and skull were in no immediate danger of ero- sion. The entire collecting process required weekly, and sometimes daily, monitoring of the site over a period often years. Excavated fossil pieces, sometimes encased within the mudstone matrix, were brought to the College of Marin Biology Laboratory in Kent- field, California, for preparation. Removal of the bone from the surrounding matrix was time-con- suming due to the highly fractured condition of the fossil, its variable state of preservation, and the hardness (cementation) of the rock. Small fossil fragments were bonded with white glue or Glyptal cement. An epoxy compound proved to be most effective for the larger pieces, especially the dentaries. Final reconstruction of the entire specimen was completed in 1986. The skull and dentaries were not assembled as single units, but as closely fitting blocks or sec- tions that can be easily disassembled for study. The assembled skull rests on a fiberglass cradle with the dentaries alongside (Fig. 2). The descriptive anatomical terminology and measurements (e.g., Tables 1-8) follow Kellogg (1922, 1928, 1931, 1934) and Miller (1923). The classification follows that of Barnes, Domning and Ray ( 1 985) and Fordyce and Barnes ( 1 994). The anatomical drawings were made by a grid and measurement method. The authors and dates of publications are provided for all cetacean taxa, and those of all fossil taxa are listed in the LIT- ERATURE Cited. The estimate of body length (10 m) was made by taking the mean of How- ell's (1930:146) formula for rorquals, based on the length of the cervical series of verte- brae. The abbreviation CASG is for the Department of Geology, California Academy of Sciences, Golden Gate Park, San Francisco, California, U. S. A. Abbreviations used in the figures are as fol- lows: apm - ascending process of maxilla be - basioccipital crest cc - coronoid crest Bo - basioccipital bone cp - coronoid process of dentary earn - channel for external auditory meatus Eoc - exoccipital bone fm - foramen magnum fn - nutrient foramina Fr - frontal bone ft - temporal fossa gf- glenoid fossa Mx - maxillary bone n - naris Na - nasal bone Oc - occipital bone occ- occipital condyle Pa - parietal bone Pal - palatine bone pgp - postglenoid process Pmx - premaxillary bone pp - posterior process of periotic sop - supraorbital process of frontal Sq - squamosal bone Tb- tympanic bulla Vo - vomer bone zyg - zygomatic process of squamosal ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 117 Figure 1 . Type locality of Parabalaenoptera baulinensis, new genus and species, indicated by arrow, at Bolinas Point, Point Reyes Peninsula, Marin County, California. Systematics Class Mammalia Linnaeus, 1758 Order Cetacea Brisson, 1762 Suborder Mysticeti Flower, 1864 Family Balaenopteridae Gray, 1864 Included Subfamilies. — Parabalaenopter- inae, new subfamily; Megapterinae (Gray, 1866) Gray, 1868; and Balaenopterinae (Gray, 1864) Brandt, 1872. Parabalaenopterinae, new subfamily Diagnosis of Subfamily. — A subfamily of the family Balaenopteridae differing from Megapterinae and Balaenopterinae by having cranium with a transversely narrower intertem- poral region; nasal bones elongate (approaching a length to width ratio of 4: 1), narrow, and nearly parallel-sided, terminating anterior to the level of the antorbital notch, rather than short, terminat- ing posterior to the antorbital notch, and distally wide; ascending process of maxilla tapered pos- teriorly rather than lobate or expanded trans- versely; exposure of frontals on dorsal surface of intertemporal region between apex of occipital shield and posterior ends of nasals, premaxillae, and maxillae nearly twice as wide anteroposteri- orly; parietals on either side of cranial vertex swollen and elevated above level of adjacent frontals; postglenoid process of squamosal short, 118 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 TABLE 1. Measurements (in mm) of cranium of holotype of P. baulinensis, new genus and species. 1. Greatest length of skull, anterior end of right premaxilla to level of posteroexternal angles of exoccipitals 2240+ 2. Distance between anterior end of right premaxilla and apex of supraoccipital shield .... 1700+ 3. Distance between posterior angles of exoccipitals to apex of supraoccipital shield 540 4. Distance from apex of supraoccipital shield to posterior end of left nasal bone 50 5. Transverse distance between outside margins of premaxillae at level of anterior ends of nasal bones 94 6. Maximum width of right premaxilla at level of anterior end of right maxilla 80 7. Transverse diameter or rostrum at level of widest point of nasal fossa 620+ 8. Greatest anteroposterior diameter of left supraorbital process 440 9. External anteroposterior diameter of left supraorbital process 210 10. Transverse diameter of cranium between outer surfaces of zygomatic processes 980 1 1. Transverse diameter of cranium between margins of exoccipitals 690 12. Width of occipital condyles 198 13. Distance from upper margin of formen magnum to apex of supraoccipital shield 455 14. Vertical distance from basioccipital to level of apex of suproccipital shield 250+ 15. Distance from anterior margin right zygomatic process to posterior margin exoccipital . . . 360 16. Distance from anterior margin left zygomatic process to posterior margin exoccipital 475 17. Greatest breadth of basioccipital between inner surfaces of notches for jugular leash 248 18. Greatest length of right palatine bone 490+ 19. Greatest width of combined palatine bones, anteriorly 275+ 20. Greatest height of left occipital condyle 131 2 1 . Greatest width of left occipital condyle 76 22. Vertical diameter of foramen magnum 93 23. Transverse diameter of foramen magnum 64 24. Greatest width of basioccipital between tympano-periotic processes 245 extending only slightly ventral to the ventral surface of the basioccipital crest; and dentary with long, sloping coronoid crest approaching coronoid process rather than an abruptly ascend- ing coronoid process; coronoid process with a spindle-shaped, posteriorly-directed apex; and anterior margin of mandibular foramen smooth, lacking notch. Parabalaenoptera, new genus DIAGNOSIS OF GENUS. — Because the sub- family Parabalaenopterinae is at present mono- typic, the diagnosis of the genus shall remain identical to that of the subfamily Parabalaenop- terinae until further genera are diagnosed in the subfamily. Type and Only Included Species. — Para- balaenoptera baulinensis, new species, Late Miocene of California. ETYMOLOGY. — From the Grtekpara, for be- side; plus balaenoptera, from Latin balaena, for whale, and Greek pteron, for fin; in reference to the apparent separation between this taxon and the lineage that includes the living balaenopter- ine whales. Parabalaenoptera baulinensis, new species (Figs. 2-7, 8b) Diagnosis of Species. — Because the genus Parabalaenoptera is at present monotypic, the diagnosis of the type species, Parabalaenoptera baulinensis, shall remain identical to that of the genus Parabalaenoptera until further species are diagnosed in the genus. HOLOTYPE. — CASG 66660, a partial associ- ated skeleton consisting of most of the cranium with tympanic bullae and periotics, both dentar- ies, fragments of the left scapula, both humeri, 23 vertebrae in varying states of preservation, in- cluding 7 cervicals, 15 thoracic vertebrae, a lum- bar vertebra, and numerous sections of incomplete ribs. ZE1GLER, CHAN, AND BARNES: PARABALAENOPTERA BAUUNENSIS 19 Figure 2. Parabalaenoptera baulinensis. new genus and species, assembled holotype cranium and dentaries, CASG 66660, oblique anterodorsolateral view. Figure 3. Parabalaenoptera baulinensis, new genus and species, oblique anterodorsal view of part of holotype skull, CASG 66660, showing elongate nasal bones. 120 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 Type Locality. — The sea cliff, above the high tide mark near Bolinas Point (37° 54' 6" N, 122° 43' 15" W), Point Reyes Peninsula, Marin County, California, U. S. A (Fig. 1). FORMATION AND AGE. --Santa Cruz Mud- stone, of latest Miocene age, between approxi- mately 6.8 and 6.0 Ma. The holotype skeleton of Parabalaenoptera baulinensis was parallel to the bedding planes in shale of the Santa Cruz Mudstone, dipping 45° to the southwest. The interbedded siltstone con- tains faint parallel laminations in beds ranging from 5 to 25 cm thick. Concretions and carbon- ate-cemented horizons are common and occur parallel to the bedding planes. Formerly consid- ered part of the Monterey Formation, these rocks are now assigned to the Santa Cruz Mudstone (Clark etal. 1984). Although not abundant, fossils such as gastro- pods, pelecypods, echinoderms, isopod frag- ments, crab claws, decapod shrimp parts, fish teeth and bones, and isolated mammal bones have been collected from this formation. Many of the fossils, including the fossil whale, are better preserved in the carbonate-cemented con- cretions, although fossils occur throughout the Santa Cruz Mudstone. Fossil mollusks in the same strata as the holotype of Parabalaenoptera baulinensis were tentatively identified as limids, possibly Acesta sp. These bivalves generally re- quire water depths of 500 m and water tempera- tures of 8°C or less (Ellen J. Moore, Branch of Paleontology and Stratigraphy, U. S. Geological Survey. Menlo Park, California, 5 May 1984. pers. comm.). An assemblage of fossil diatoms was recov- ered from carbonate concretions in the same bed- ding plane adjacent to the fossil whale. The diatom assemblage is assignable to "Subzone A" of the Late Miocene Nitzschia reinholdi Zone, and implies an estimated age of 6.8 to 6.0 Ma for the fossil whale (John A. Barron. Branch of Paleontology and Stratigraphy, U. S. Geological Survey, Menlo Park, California, 10 January 1985, pers. comm.). ETYMOLOGY. - The species name, baulinen- sis, is in reference to the type locality at Bolinas Point, Point Reyes Peninsula, Marin County, California; Bolinas is named after and was part of the 1860 historical land title of Rancho las Baulines. Description. — Cranium. — Characters that distinguish the cranium of Parabalaenop- tera baulinensis from those of Balaenoptera Lacepede, 1 804, and Megaptera Gray, 1 846, are: the more narrow intertemporal region (Figs. 4, 6), the long and narrow nasals (Figs. 3, 4, 6), and the laterally swollen and elevated portions of the parietals on either side of the intertemporal re- gion (Fig. 7b). The entire rostrum is not preserved and some large sections are missing. In order to estimate the original total length of the skull, we calcu- lated the ratios of rostral length to total skull length in large extant species of rorquals. The average of these gives a rostral length that is approximately 74% of the total skull length. We applied this ratio to the fossil to obtain its esti- mated original skull length and body length. The rostrum is very flat and gradually in- creases in depth posteriorly. The lateral margin of the maxilla is thin and blade-like. The maxilla thickens medially as it curves medially and ven- trally to enclose the vomer. The distal ends of the premaxillae are missing anterior to the distal ends of the maxillae (Fig. 6). The portion of the left premaxilla anterior to the narial opening has been pressed ventral to and against the right premaxilla as a result of geologi- cally induced right-lateral compression of the cranium. In life, these bones were definitely separated, and probably somewhat parallel, up to their point of divergence at the anterior margin of the vertical orientation as they delineate the long, broad narial fossa. Lateral to the nasals, the premaxillae become progressively narrower pos- teriorly, only being 10 mm wide at the posterior ends of the nasals, and terminate 45 mm from the apex of the supraoccipital. The maxillae are fractured and are missing some portions on their lateral margins (Fig. 6). What is preserved of the rostrum indicates that it was very broad, not arched anteroposteriorly, and that its lateral margins were thin and curved ventrally. The posterior parts of the rostral mar- gins are fragmentary and their exact relationship to the antorbital processes is difficult to deter- mine. The long, ascending (posterior) processes of the maxillae, overlapping and penetrating the frontals, are cuneate in shape, tapering gradually posteriorly to within 30 mm of the anterior mar- gin of the supraoccipital, and are each 15 mm ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 121 fracture Figure 4. Parabalaenoptera baulinensis, new genus and species, dorsal view of holotype cranium and dentaries, CASG 66660; fractured area of left dentary, discussed in text, is indicated. I L 200 mm Figure 5. Parabalaenoptera baulinensis, new genus and species, lateral view of holotype cranium and right dentary, CASG 66660; see Materials and Methods for abbreviations. wide at the level of the posterior margin of the nasals. The nasals (see Figs. 3, 6) are distinctive by being extremely long (maximum length = 250 mm) and narrow for a balaenopterid, and this is a primitive character. Their combined anterior width, as preserved now is 57 mm, but we esti- mate that this width was 64 mm prior to geologic distortion. The combined posterior transverse width of the nasals is 29 mm. The nasal bones are tightly wedged between the ascending processes of the premaxillae, and terminate 50 mm from the apex of the supraoccipital. The length-to-anterior width ratio of the com- bined nasals is approximately 4:1 (Fig. 3). By contrast, among the living balaenopterids, this ratio ranges from approximately 2:1 in Balaenoptera borealis Lesson, 1828, (Sei whale; see Fig. 8c), to approximately 1:1 in Balaenop- tera physcdus (Linnaeus, 1758) (fin whale; see Fig. 8d), which has exceptionally short nasals (a derived character). The anterior ends of the nasals in Parabalaenoptera baulinensis extend anteriorly farther than the anterior margin of the supraorbital processes. Using the supraorbital process as a point of reference for comparison with living balaenopterids, only in the blue whale 122 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 300 mm Figure 6. Parabalaenoptera baulinensis, new genus and species, dorsal view of holotype cranium, CASG 66660; see Materials and Methods for abbreviations. ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 123 Eoc 300 i mm occ Figure 7. Parabalaenoptera baulinensis. new genus and species, braincase of holotype, CASG 66660; a. posterior view; b. lateral view; c. ventral view; see Materials and Methods for abbreviations. 124 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 Balaenoptera musculus (Linnaeus, 1758); see Fig. 8g) do the nasal bones extend more pos- teriad. Concomitant with the elongate nasals is the conspicuously anterior position of the narial opening, which has its maximum width of 140 mm at a point 180 mm anterior to the ante- rior ends of the nasals. The dorsoventral depth of the narial opening at this point, from the dorsal surfaces of premaxillae to the vomer, is 135 mm. The anteroposterior diameter of the narial open- ing, taken from the point of divergence of the premaxillae to the anterior ends of the nasals, is 520 mm. The choanae, ventral to the nasals, are filled with matrix and not visible for inspection. Both supraorbital processes of the frontals are fragmented, especially the right process, which crumbled during excavation; however, the left supraorbital process is nearly complete (Fig. 6). Unfortunately, the relationship of the antorbital process to the lacrimal bone cannot be deter- mined as the lacrimal bones are both missing. As is typical of balaenopterids, the supraorbital process is broad, somewhat flat, approximately rectangular in dorsal view, and its dorsal surface slopes ventrally away from the higher cranial vertex. The supraorbital process is broad antero- posteriorly at its medial part. The lateral orbital margin is moderately arched dorsally. The flat plate of the supraorbital process is very long anteroposteriorly at the base and is abruptly de- pressed well below the dorsal level of the interor- bital region. The lateral margin of the supraorbital process is slightly arched over the orbit, and the postorbital process is small as in Recent balaenopterids. The optic nerve canal is enclosed proximally, but as it proceeds laterally, the canal opens and flares out to form the slightly arched roof of the orbit. In the intertemporal region the parietals are swollen on either side of the frontals and con- spicuously raised laterally above the nearly ver- tical lateral margins of the interorbital region. An anterior projection of the anteroexternal portion of the parietals extends over the frontals on either side of the intertemporal region and the cranial vertex, and they extend on either side of the intertemporal region anteriorly beyond the pos- terior ends of the nasals, the premaxillae, and ascending processes of the maxillae. The cranial vertex has the typical configuration of a balaenopterid, with the rostral bones deeply interdigitating with the frontals. There is a small, somewhat rectangular exposure of the frontals on the vertex, but it is still larger than in any Recent balaenopterid (Fig. 8). The narrow (for a balaenopterid) interorbital part of the frontals is exposed between the posterior ends of the nasals, the maxillae, and the premaxillae. The triangular occipital shield tapers to a trun- cated apex, which extends anteriorly to a point beyond the level of the anterior extremities of the zygomatic processes of the squamosals, ap- proaching a line drawn transversely through the centers of the orbits. The anterior extent of the supraoccipital shield and the parietals in P. baulinensis exceeds that of B. musculus, and closely approaches that of the other Recent spe- cies of Balaenopteridae. The apex of the occipital shield is not clearly defined, due to apparent abrasion prior to deposition. The essentially straight lateral margins of the occipital shield project laterally over the temporal fossae. There is no concavity to the lateral margin of the occipi- tal shield such as is present in all living species of Balaenoptera (Fig. 8). The mid-part of the occipital shield is depressed, attaining a maxi- mum depression of 85 mm below the lateral mar- gins. A lambdoidal crest, formed by the conjunction of the lateral margins of the supraoc- cipital with the posterior margins of the squamosals and the parietals, arcs dorsomedially toward the truncated apex of the supraoccipital shield. The braincase of P. baulinensis is shallow dorsoventral ly (Fig. 5), notwithstanding the geo- logic compression of the skull. The wide exoc- cipitals flare ventrolateral ly (Fig. 7a). The exoccipitals protrude only slightly beyond the posterior margin of the squamosal, and comprise the posterolateral corners of the cranium. The parietal suture with the squamosal is ob- scure, however, where the parietal overrides the supraorbital process of the frontal, the bounda- ries of this bone can be traced. The anterior-most extension of the parietal lies 100 mm distal to the posterior level of the ascending process of the maxilla. The zygomatic process of the squamosal is large, excavated ventrally, and arches upward at a steep angle from the shallow postglenoid proc- ess. It terminates in a stout, anteromedially- directed anterior tip. The occipital condyles are widely separated dorsally, but are closely approximated ventrally. ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 125 b f g h Figure 8. Dorsal views of skulls of fossil and Recent Balaenopteridae; a. Plesiocetus cortesii (after Van Beneden, 1 875:fig. 1); b. Parabalaenoptera baulinensis, new genus and species; c. Balaenoptera borealis (after Miller, 1924b:pl. 1); d. Balaenoptera physalus (after True, I904:pl. 1, fig. 3); e. Balaenoptera ede/?/ (after Omura, 1959:pl. l,fig. \);f. Balaenoptera acutorostrata (after True, 1 904:pl. 22, fig. 1 ); g. Balaenoptera musculus (after Miller, 1 924a:pl. 1 ); h. Megaplera novaeangliae (after True, 1904:pl. 29); not to scale. Their articular surfaces are narrowest dorsally and broadest near the ventral margin of the fora- men magnum. There is a slight constriction or neck around the bases of the occipital condyles, which have slightly rugose articular surfaces. The foramen magnum is posterodorsally ori- ented. The flat ventral surfaces of the basioccipital crests are on the same plane as the ventral mar- gins of the exoccipitals. The postglenoid proc- esses (these are wide but very shallow dorsoven- trally), extend only about 30 mm ventrally. Although there is some evidence of minor abra- sion of the ventral surface of the postglenoid processes, it is not enough to affect their dor- soventral dimension. A vertically oriented slot-like foramen pseudovale (37 mm x 13 mm) is open anterolat- 126 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 erally as in Recent rorquals. It is located in the squamosal immediately above the ventral junc- ture with the pterygoid at about the same level as the anterior margin of the scaphoid fossa. The ventral surface of the rostrum is not as complete nor as well preserved as the dorsal surface. The infraorbital processes of the maxil- lae, characteristic of all mysticetes, have been broken off and are missing. The boundaries of the palatines can only be barely discerned, but their anterior and lateral margins can be faintly traced. Their posterior margins are obscure, except for the divergent medial margins that reveal a shal- low vomer (20 mm deep at this point), which forms the vertical septum between the narial passages (Fig. 7c). The relatively small cross- sectional area of the nasal choanae is a primitive condition, and may indicate a less efficient div- ing ability compared to the living rorquals. The exposed vomer septum, extending posteriorly on the basicranium, decreases in height posteriorly until it disappears approximately 220 mm from the posteromedial margins of the palatines. The distributions of some of the basicranial bones are unclear because of fusion of the sutures between them and the incomplete preservation. The basioccipital is transversely concave be- tween the descending basioccipital crests, which have relatively flat ventral surfaces and have a dorsoventral dimension of 57 mm. The greatest transverse distance between the medial margins of the ventral surfaces of the basioccipital crests is 1 16 mm. Both tympanic bullae were recovered and the left bulla remains in place in the basicranium. The right bulla measures 80 mm anteroposteri- orly, 52 mm transversely, and 41 mm in depth. The tympanic bullae extend approximately 10 mm below the basioccipital crests (Fig. 7a). The relatively long posterior processes of the peri- otics are wedged between the squamosals and exoccipitals (Fig. 7c). The postglenoid process is broad transversely and very shallow dorsoventral ly. The slightly posteriorly directed postglenoid process ob- scures (in ventral view) the central part of the deep external acoustic meatus. This meatus is straight and is almost at 90° to the sagittal plane. The zygomatic process of the squamosal projects laterally and anteriorly about 45° to the sagittal plane of the cranium. It arches dorsally, terminat- ing in an enlarged, medially directed anterior extremity. The occipital condyles do not extend so far posteriorly as do the posterolateral corners of the exoccipitals. The exoccipitals are very com- pressed anteroposteriorly, being not much thicker in this dimension than the lateral end of the posterior process of the periotic. Dentary. — Both dentaries (Figs. 4, 5) are nearly complete, except for a ± 100 mm section of the right horizontal ramus about 120 mm from the anterior tip, and a few small fragments miss- ing from the coronoid processes and the dorsal margins along the dental foramina (see Table 2). There is evidence that the left dentary was broken in life. An apparent fracture (Fig. 4) is located about 550 mm from the anterior end. It appears that the break was at least partially healed, leaving a ± 25 mm offset of the anterior portion. Subsequent transverse fracturing of the fossilized bone in this area is unclear. The injury TABLE 2. Measurements (in mm) of right dentary of holotype of Parabalaenoptera baulinensis, new genus and species. 1 . Greatest length of dentary along outside curvature 2340± 2. Greatest length of dentary in a straight line 2240+ 3. Greatest vertical diameter of dentary at level of symphysis 74 4. Greatest transverse diameter of dentary at level of symphysis 47 5. Greatest vertical diameter of dentary at a point 850 mm anterior to hinder face of condyle. 150 6. Greatest vertical diameter of dentary at coronoid process 124 7. Least vertical diameter of dentary behind coronoid process 245 8. Greatest vertical diameter of condyle 180 9. Greatest transverse diameter of condyle 127 10. Greatest dorsoventral diameter of portal to mandibular canal 66 1 1 . Greatest transverse diameter of portal to mandibular canal 43 ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENS1S 127 to the left dentary might account for its being more slender than the right dentary. Seven nutrient foramina (Figs. 4, 5) are located along the dorsolateral side of each dentary. The spacing between these foramina varies between the two dentaries. The dentary is conspicuously bowed laterally (Figs. 4, 5), with no evidence of torsion or lateral flattening of the anterior portion. It increases in transverse thickness very gradually to about the level of the posterior margin of the coronoid process, where it then narrows to become con- stricted between the coronoid process and the mandibular condyle. In addition to the difference in thickness be- tween the two dentaries, the coronoid regions of the two dentaries show some differences. The right coronoid process is somewhat longer an- teroposteriorly, and has a greater lateral curva- ture than does the left. This difference in curvature is probably due to distortion pressure from the left that straightened the left coronoid, while exaggerating the curvature of the right coronoid (Fig. 5). If this is the case, the distortion was very uniformly distributed, because the out- ward arc of the right coronoid process is smooth throughout its entire length. At a point on the dentary about two-thirds of its length from the anterior tip, the coronoid crest begins its long, sloping elevation from the dorsal margin of the ramus. As the dorsal margin of the coronoid crest sweeps gradually upward, it also begins arcing laterally. At its apex, the coronoid process forms a pointed and posteriorly directed tip creating a boss (Fig. 5). The fossa that is confluent with the mandibular foramen begins near the mandibular condyle, and is directed slightly ventrally, and deepens as it progresses anteriorly. At a point just posterior to the coronoid process, the mandibular foramen enters the dentary with an arch-like margin curv- ing smoothly ventrally and slightly posteriorly; consequently, there is no notch in the anterior margin of the foramen as in Balaenoptera and Megaptera. Vertebrae. — The preserved anterior part of the vertebral column includes the seven cervi- cals, fifteen thoracic vertebrae in series, and an isolated lumbar vertebra. The centra of these vertebrae are mostly well preserved and show little evidence of abrasion. The centra of vertebrae in the thoracic and lumbar regions are uniformly separated by ma- trix that filled the inter- vertebral spaces. The neural arches are only fragmentary and the trans- verse processes have been broken away from the centra of most of the vertebrae. Cervicals 6 and 7, and the first and second thoracic vertebrae remain in a block of concretionary matrix. The vertebral epiphyses are completely fused to all of the centra, indicating that the holotype was an adult individual. Atlas. — The neural arch of the atlas (see Ta- ble 3) is missing, as is the right parapophysis. The posterior articular surfaces are convex. Along the ventral margin of the atlas is a lip-like projection, most prominent at the midline, which extends posteriorly a short distance under the ventral surface of the axis, and conforms closely to the contours of the latter. The imperforate left trans- TABLE 3. Measurements (in mm) of atlas of holotype of Parabalaenoptera baulinensis, new genus and species. Greatest breadth across transverse processes 354+ Distance ventral margin midline to dorsal margin adontoid fossa 50 Greatest height of right anterior articular surface 150 Greatest breadth of right anterior articular surface 85 Greatest breadth of left anterior articular surface 75 Greatest transverse diameter of neural canal 80 Maximum breadth across posterior articular surfaces 218 Greatest height of right posterior articular surface Ill Greatest height of left posterior articular surface Ill Greatest breadth of right posterior articular surface 80 Greatest breadth of left posterior articular surface 80 Greatest anteroposterior diameter 68 Maximum anteroposterior diameter at ventral midline 57 128 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 TABLE 4. Measurements (in mm) of axis of holotype of Parabalaenoptera baulinensis, new genus and species. 1. Greatest breadth across transverse processes 500± 2. Height of right transverse process at exterior margin of transverse foramen 143 3. Height of left transverse process at exterior margin of transverse foramen 130 4. Transverse diameter of right transverse foramen 67 5. Transverse diameter of left transverse foramen 62 6. Dorso-ventral diameter of right transverse foramen 38 7. Dorso-ventral diameter of left transverse foramen 32 8. Distance of external margin right articular surface to internal margin right transverse foramen 12 9. Distance of external margin left articular surface to internal margin right transverse foramen 15 10. Greatest breadth of anterior articular surfaces 204 1 1. Height of right anterior articular surface 121 12. Height of left anterior articular surface 120 13. Maximum extension of odontoid process above the posteriormost level or articular surface . 18 14. Distance of anteriormost margin of odontoid process above midline ventral margin 50 15. Greatest posterior transverse diameter of centrum 160 16. Greatest posterior dorsoventral diameter of centrum Ill 17. Greatest anteroposterior diameter of centrum at dorsal level 50 18. Greatest anteroposterior diameter of centrum at ventral level 37+ verse process is short and stout, tapering to a rounded end. This process extends 55 mm be- yond the lateral margin of the left posterior ar- ticular surface at a level slightly below the dorsal margin of this facet. Axis. — The neural arch of the axis (Table 4) and a small portion of the right transverse process are missing. There is a rudimentary odontoid process projecting anteriorly only 18 mm from the concave anterior articulating surfaces. The long and deep, wing-like, transverse processes are posteriorly deflected approximately 30° from the transverse plane of the centrum. Third to Seventh Cervical Vertebrae. — The third cervical (see Table 5) is a very fragile vertebra compared to the robust axis. The supe- rior and inferior transverse processes meet to delineate the relatively large transverse foramen. The left foramen has an estimated transverse diameter of 100 mm and a dorsoventral diameter of 75 mm. The neural canal is 95 mm in trans- verse diameter. The greatest breadth across the transverse processes would appear to approach that of the axis. The fourth cervical is the most anteroposteri- orly compressed cervical vertebra. Its neural ca- nal has an estimated dorsoventral diameter of 50 mm and an estimated transverse diameter of 95 mm. The fifth cervical vertebra reverses the trend of anteroposterior compression from the atlas to- ward cervical 4, with the sixth and seventh cerv- icals becoming progressively thicker antero- posteriorly. The neural arches are crushed upon the centra of all three of these vertebrae. Thoracic and Lumbar Vertebrae. — Due to the fragmentary condition or absence of trans- verse processes on the series of 15 vertebrae posterior to the cervicals, it is not possible to determine the transition point from the thoracic to lumbar vertebrae. It is possible that the series are all thoracic vertebrae, inasmuch as the num- bers of thoracic vertebrae in balaenopterid whales range from 1 1 to 15 (Kellogg 1925). With this variability in mind, we refer to the vertebrae as the thoracic-lumbar series (Table 6); the more complete twelfth vertebra in that series is treated more extensively (Table 7). The first thoracic has a centrum that is larger dorsoventral ly than any others in the thoracic-lumbar series, 3 through 15, and the width of the anterior face of its centrum is exceeded only slightly by the second thoracic. There is a progressive increase in length of the centra posteriorly in the column. The iso- lated lumbar vertebra is smaller in all dimensions compared with the last vertebra in the preceding series. ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 129 TABLE 5. Measurements (in mm) of cervical vertebrae 3-7 of holotype of Parabalaenoptera baulinensis, new genus and species. Cervical vertebrae C-3 C-4 C-5 C-6 C-7 1 . Length of centrum at dorsal level 2 . Height of anterior face of centrum 3 . Width of anterior face of centrum 4. Length of centrum at ventral level Ribs. - The ribs are very fragmentary, but parts of eight right ribs and one left rib have been assembled. The right rib that was found adjacent to the right humerus measures 820 mm in length along its exterior margin. The narrowest part of this rib is at the neck, where it measures 26 mm anteroposteriorly, and the thickest part of this rib in this dimension is 43 mm, midway between the tuberosity and the exterior margin of the angle. In anterior view the rib is widest (80 mm) at the angle, and the transverse diameter at the broken distal portion is 68 mm. Scapula. — Only fragments of the left scapula were collected. The largest of these is the distal portion of the scapula bearing the glenoid area. The glenoid fossa has a maximum anteroposte- rior diameter of 1 50 mm. There is a protuberance with a broken surface about 20 mm above the anterior margin of the glenoid fossa that may be the base of the coronoid process; it is directed anteriorly and slightly medially. Humerus. — Both humeri are relatively intact and are fairly complete (Table 8). The margins of the tuberosities and distal portions are eroded and a small central section of the left shaft is missing. The large, globular head is smooth and connected by a short but prominent neck to the shaft, which is quite flat mediolaterally. The dis- tal end has two trochlear facets, which are slightly concave, for articulation with the radius and ulna. On the right humerus, the radial facet is at an angle of approximately 48° to the plane of the ulnar facet. Erosion has made it difficult to estimate the dimensions of the trochlea. 41 105 157 38 35 123 150 34 39 155 40 150 163 41 150± 160± 47 Discussion Balaenopterid Taxonomy and the Problem of Disparate Parts Not all named fossil balaenopterids are based on directly comparable elements. Regarding mysticete taxonomy, Kellogg (1931) has stated that the classification of fossil taxa must be based upon cranial characters, an opinion that was shared by Barnes (1976). For our comparisons, we examined 14 fossil species of mysticetes in 14 genera. Nine of these species are represented by parts of the cranium, and three of these also include parts of dentaries. The remaining five species are represented by only a single dentary or by parts of dentaries. In contrast, the holotype of P. baulinensis is represented by about 90% of the cranium and dentaries, along with both humeri and 23 vertebrae, making it one of the most complete fossil mysticete specimens yet described. Because of its completeness, it was possible to compare it to and distinguish it from all other relevant mysticete taxa. Other Described Fossil Balaenopterids Fossil balaenopterids are relatively abundant in late Cenozoic marine deposits around the world (Barnes 1976). In publications, some of these fossils have been assigned to living genera, but have not been more precisely identified. Most of these less precisely identified fossils are clearly referable to the subfamilies Balaenopter- inae and Megapterinae and are not directly rele- vant to the present study. It is, however, necessary to consider previously named balaenopterids in arguing that Parabalaenoptera baulinensis is a previously unrecognized taxon in anew subfamily. The extinct balaenopterids that were described by Van Beneden (e.g., 1872, 1880, 1882, 1885) 130 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 on o VI U ■c 4> <*! CO O CO O E £ c — 3 on cO nO W J < CM On OO r- nO cm fN nO OO I/O CM IT> o no CO co O CO O CO 10 in no CO «0 o 00 ON o r- co CO ON CM OO CM On CM ir> CM 00 CM O CO o CO O CO CM CM O NO CM NO r- o r- 10 o NO o NO o NO -3- NO 00 10 CM uo CO I/O CO CM CO CO MHO CM o © CM OO ON r- 00 00 >/o CM NO ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 131 Table 7. Measurements (in mm) of 12th vertebrae of dorsal-lumbar of holotype of Parabalaenop- tera baulinenis, new genus and species. Distance from front of anterior zygapophysis to posteriormost margin of neural spine -190 Distance from level of ventral margin anterior face of centrum to dorsal level of anterior zygapophysis 205 Greatest dorsoventral diameter of neural canal at anterior dorsal margin of centrum 60 Greatest transverse diameter of neural canal 55 Greatest length of neural canal 100 Greatest width of neural spine at dorsal base of anterior zygopophysis 27 Length of base of right transverse process 115 Height of base of right transverse process 34 TABLE 8. Measurements (in mm) of humeri of holotype of Parabalaenoptera baulinensis, new genus and species. 1. Greatest length 2. Greatest width posterior margin of shaft to anterior margin of greater tuberosity 3. Anteroposterior diameter of distal end 4. Exterointernal diameter of distal end 5. Anteroposterior diameter of head 6. Proximal-distal diameter of head Right Left 305 — 135 — -130 — 67 132 130 101 115 from the Antwerp Basin in Belgium belong to the closely related balaenopterine genera Balaenop- tera and Burtinopsis, to the megapterine genus Megaptera, and to the primitive balaenopterid genus Plesiocetus Van Beneden, 1859. Most of the species that Van Beneden described are based on rather complete associated skeletons, but in most cases the specimens include only parts of the crania and mandibles. Included among the Belgian fossils are the extinct balaenopterines Balaenoptera musculoides Van Beneden. 1880; Balaenoptera borealina Van Beneden, 1880; and Balaenoptera rostratella Van Beneden, 1 880; and the closely related species Burtinopsis similis Van Beneden, 1872, and Burtinopsis mi- nutus Van Beneden, 1880. Balaenoptera muscu- loides and B. borealina both have parallel dorsal and ventral margins of the horizontal ramus of the dentary, without the long ascending coronoid crest that characterizes Parabalaenoptera baulinensis, and in this regard they are like the living species of Balaenoptera. Likewise, B. ros- tratella has characters of the genus Balaenop- tera, and is approximately one-half the size of Parabalaenoptera baulinensis. Species of Burti- nopsis are similar to Balaenoptera, and in fact. Demere (1986) has considered Burtinopsis to be a junior synonym of Balaenoptera. Because the known morphologies of these fos- sil balaenopterids are consistent with those of the living species of Balaenoptera and Megaptera, we accept Van Beneden's original generic allocations for them and assign them to the sub- families Balaenopterinae and Megapterinae, re- spectively. Each of them can be distinguished from Parabalaenoptera by the characters that we use to diagnose the subfamily Parabalaenopteri- nae (Table 9). A species in the supposed balaenopterid genus Mizuhoptera Hatai, Hayasaka, and Masuda, 1963, was described on the basis of isolated tympanic bullae of Pliocene age from Japan. Oishi and Hasegawa (1995) concluded that the type material actually represents the cetotheriid genus Herpetocetus Van Beneden, 1872. There- fore, the genus Mizuhoptera is invalid and is not a balaenopterid. A Late Miocene mysticete, Balaenoptera Ir- ani Hanna and McLellan, 1924, from the Mon- terey Formation in central California, was named 132 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 TABLE 9. Differences between Parabalaenopterinae, new subfamily, and the subfamilies of Megapterinae and Balaenopterinae. Parabalaenopterinae Megapterinae and Balaenopterinae CRANIUM 1. Cranial vertex 2. Nasals 3. Shape of ascending process of maxilla 4. Anterior portion of parietal 5. Postglenoid process DENTARY 6. Length of coronoid ridge anterio- posteriorly relative to mandible length 7. Direction of terminus of coronoid process 8. Terminus of coronoid process 9. Mandibular foramen narrow long, narrow cuneate raised, swollen shallow long posterior bullet-shaped boss lacks notch broad short, broad flared flat, feathered deep and moderately deep short dorsal smoothly tapered notched on the basis of a braincase. It is not a member of the genus Balaenoptera, and is not even a balaenopterid. Instead, it is a typical cetotheriid (see Barnes, 1976:330) because its parietals are exposed in a wide band across the top of the intertemporal region, and it does not have the type of interfingering of the rostral and cranial bones that is diagnostic of the family Balaenop- teridae. Only two fossil species of balaenopterids have been previously named from the Pacific coast of North America. Megaptera miocaena Kellogg, 1922, from the Late Miocene Monterey Forma- tion in Santa Barbara County, California, is a large species of balaenopterid (Barnes, 1976:329). It has short and wide nasal bones and a wide intertemporal region like both Megaptera and Balaenoptera, but the concave lateral mar- gins of its occipital shield are like members of the subfamily Balaenopterinae. Its zygomatic arches are relatively narrow, a primitive character, and it appears to be a primitive species of the subfam- ily Balaenopterinae. Balaenoptera davidsonii Cope, 1872, is a small mysticete from the Late Pliocene age San Diego Formation at San Diego, California. It was first named as a species of Eschrichtius (gray whales), but Barnes (1976:333, table 6) removed it from the Eschrichtiidae, pointing out that the holotype dentary did not have characters of the family, and Demere( 1986) placed it in the genus Balaenoptera. Thus, both previously named fos- sil balaenopterids from the eastern North Pacific realm are apparently members of the subfamily Balaenopterinae. The Relationships of Parabalaenoptera baulinensis Parabalaenoptera baulinensis is demonstra- bly a member of the family Balaenopteridae, which includes the living humpback {Megaptera novaeangliae (Borowski, 178 1 )), various species of rorquals of the genus Balaenoptera, and vari- ous extinct taxa as mentioned above. Para- balaenoptera baulinensis possesses six derived cranial characters and one derived mandibular character that are diagnostic of the family Balaenopteridae (see also Miller 1923; Kellogg 1931). These derived characters, which define the family Balaenopteridae, are listed in the leg- end (Node 1) of the cladogram (Fig. 9), and are discussed as follows: 1) The parietals are excluded from medial ex- posure on the dorsal surface of the intertemporal region, but they extend anteriorly along either side of the frontals at least to a point beyond the posterior ends of the rostral bones (the nasals, maxillae, and premaxillae). 2) The apex of the occipital shield extends anteriorly at least as far as the level of the poste- ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 133 rior border of the supraorbital process and the anterior extremity of the zygomatic process of the squamosal. 3) The frontal has only a small exposure on the middle of the dorsal surface of the interorbital region. 4) The lateral part of the squamosal is rela- tively large and expanded, and the postglenoid process is large. 5) The dorsal surface of the supraorbital proc- ess of the frontal is very depressed medially, descending abruptly from the more elevated dor- sal surface of the interorbital region. 6) The maxillary part of the rostrum is trans- versely expanded and relatively flat, except in its median part, which has a keel that is created by the vomer, and is not conspicuously arched an- teroposteriorly. 7) The dentary is conspicuously bowed later- ally, beyond the curvature of the rostral margin. Parabalaenoptera baulinensis has five primi- tive characters that exclude it from the derived subfamilies Balaenopterinae and Megapterinae, and these contribute to the definition of the new subfamily Parabalaenopterinae. These five char- acters are discussed as follows: 1) The narrow intertemporal region of Para- balaenoptera baulinensis is a retained primitive character that is shared with members of the family Cetotheriidae and with the primitive balaenopterid Plesiocetus cortesii (Capellini, 1865) (see discussion of Plesiocetus below). Members of the Balaenopterinae and Megapter- inae (see Kellogg 1928; Barnes and McLeod 1984) have relatively wider intertemporal re- gions than P. baulinensis, in some cases wider by a factor of more than two. 2) The dorsal exposure of the frontals on the middle of the dorsal surface of the intertemporal region between the posterior ends of the rostral bones and the apex of the occipital shield in Parabalaenoptera baulinensis is relatively large for a balaenopterid, forming an approximately rectangular patch. This larger exposure is a primitive character that is shared with members of the family Cetotheriidae, and contrasts with the much reduced and nearly absent dorsal expo- sure of the frontals in skulls of animals in the subfamilies Balaenopterinae and Megapterinae. 3) The nasals of Parabalaenoptera baulinensis are longer (4: 1 length to width ratio) than those of any extant balaenopterid, extending anteriorly to a point anterior to the level of the antorbital notches. In members of the subfamilies Balaenopterinae and Megapterinae, however, the nasals are short and terminate posterior to the level of the antorbital notches. The nasals of P. baulinensis are nearly as long as those of such Late Oligocene and Early Miocene cetotheriids (sensu Kellogg 1931) as Mauicetus parki (Ben- ham, 1937) (nasal length to width ratio of 5.5:1) and Aglaocetus moreni (Lydekker, 1894) (nasal length to width ratio of 5: 1 ). Therefore, the long and narrow nasals of Parabalaenoptera baulinensis are a retained primitive character. 4) The posteriorly tapering, narrow, ascending processes of the maxillae of P. baulinensis differ from the lobate posterior ends of the maxillae in the Balaenopterinae and Megapterinae. In P. baulinensis these processes become progres- sively more narrow posteriorly and terminate in rounded points. This condition we refer to as "cuneate," in order to distinguish it from the transversely flared or lobate shape of the ascend- ing processes, which was considered by Barnes and McLeod (1984) to be a balaenopterid char- acter. Henceforth, the posteriorly expanded as- cending processes of the maxillae shall be considered a derived character that is shared by the subfamilies Balaenopterinae and Megapteri- nae. The cuneate processes are a primitive char- acter of Parabalaenoptera baulinensis, which is shared with cetotheriids. and distinguishes the subfamily Parabalaenopterinae from the two other balaenopterid subfamilies. 5) The short postglenoid process of Para- balaenoptera baulinensis is a primitive character that is shared with species of Cetotheriidae and primitive balaenopterids of the genus Plesio- cetus. It differs from the bulbous postglenoid process of the derived balaenopterids, and is a defining character of the subfamily Para- balaenopterinae. The derived character states of the above five structures (the opposite polarity of the character states in Parabalaenoptera baulinensis) are shared by the subfamilies Balaenopterinae and Megapterinae, and indicate that the latter two are sister taxa. They appear at Node 5 on the cladogram (Fig. 9). Parabalaenoptera baulinensis has three auta- pomorphies, which appear at Node 4 in the 134 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 Cetotheriidae Plesiocetus Parabalaenoptera Balaenopterinae Megapterinae cortesii baulinensis Figure 9. Cladogram showing postulated relationships of Parabalaenoptera baulinensis, new genus and species, among fossil and Recent Balaenopteridae. The subfamily Cetotheriinae is the outgroup. Characters marking the numbered dichotomies and end points are as follows: 1 . Anterior border of supraorbital process of frontal depressed relative to dorsal surface of maxilla and forming a coved recess; parietals excluded from dorsal surface of intertemporal region and extending anteriorly on lateral sides of the frontals in the intertemporal region to a point beyond the posterior ends of rostral bones; apex of occipital shield extended anteriorly at least as far as the level of the posterior border of the supraorbital process and the anterior extremity of the zygomatic process of the squamosal; lateral part of the squamosal relatively large and expanded; supraorbital process of the frontal abruptly depressed at its medial part to a level well below that of the dorsal surface of the interorbital region; dentary conspicuously bowed laterally [family Balaenopteridae]. 2. Nasals short, terminating anteriorly at a point posterior to the level of the antorbital notch [Plesiocetus cortesii]. 3. Contact between the posterior border of the maxilla and the anterior margin of the supraorbital process of the frontal curved anteriorly; postorbital process of frontal reduced; zygomatic process of squamosal inflated and curved medially (concomitant with shorter postorbital process); mandibular condyle spherical and bulbous. 4. Parietal on either side of the intertemporal region swollen and elevated dorsal to the level of the adjacent frontal; coronoid crest of dentary elevated and inclined toward the coronoid process; apex of coronoid process posteriorly directed and spindle-shaped [Parabalaenoptera baulinensis]. 5. Nasal bones short and terminating posterior to the antorbital notch; posterior end of the ascending process of the maxilla lobate or transversely expanded; intertemporal region expanded transversely; dorsal exposure of the frontals between posterior ends of the rostral bones and apex of the occipital shield reduced; parietals on either side of intertemporal region thin and interfingering with frontals anteriorly; postglenoid process of the squamosal large, bulbous, and extending ventrally from the cranium. 6. Concave lateral margin of the occipital shield where it overhangs the temporal fossa [subfamily Balaenopterinae]. 7. Braincase transversely expanded; coronoid process of dentary reduced in size; olecranon process of ulna lost [subfamily Megapterinae]. cladogram (Fig. 9). These contribute to the defi- nition of the taxon, and are discussed as follows: 1 ) In Parabalaenoptera baulinensis, as in all Balaenopteridae, the parietals are excluded from exposure in the dorsal surface of the intertempo- ral region, and they extend anteriorly along both sides of the frontals to a point as far as the posterior ends of the nasals. However, a unique derived character of P. baulinensis is the infla- tion and elevation of this lateral part of the parie- tals above the level of the frontals. Both the right and left parietals exhibit this feature to about the same degree, indicating that it is normal and not pathological. This contrasts with the flat, inter- fingering condition of the parietals in the sub- families Balaenopterinae and Megapterinae. 2) The long, gradually ascending coronoid crest on the dentary that approaches the coronoid process (see Fig. 5) is a unique derived character of P. baulinensis. The structure is absent in all other Balaenopteridae, in which the coronoid process rises abruptly from the straight dorsal ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 135 other Balaenopteridae, in which the coronoid process rises abruptly from the straight dorsal margin of the horizontal ramus. 3) The apex of the coronoid process is posteri- orly directed and spindle shaped. This is unlike the condition in Cetotheriidae and in the Balaenopterinae and Megapterinae. To summarize the points above, Para- balaenoptera baulinensis is a member of the family Balaenopteridae, because it has the suite of diagnostic osteological characters that are presently used to recognize the family. It also has some plesiomorphic characters that are not pre- sent in members of the two highly evolved balaenopterid subfamilies. Balaenopterinae and Megapterinae: narrow intertemporal region, larger exposure of frontal in intertemporal re- gion, long and narrow nasal bones, tapered as- cending processes of the maxillae, and short postglenoid process of the squamosal. These primitive characters are shared with members of the family Cetotheriidae, and they add to evi- dence for derivation of the Balaenopteridae from the Cetotheriidae. Parabalaenoptera baulinensis also has three unique derived characters that separate it from other Balaenopteridae: the ele- vated and swollen parietals on either side of the intertemporal region, the long, ascending coronoid crest anterior to the coronoid process on the dentary, and the spindle-shaped posteriorly- directed apex of the coronoid process. The com- bination of the five primitive characters and three unique derived characters of Parabalaenoptera baulinensis serve to define the taxon and the new subfamily Parabalaenopterinae that contains it. The Genus Plesiocetus Van Beneden, 1859 The family assignment of the extinct mysticete genus Plesiocetus Van Beneden, 1 859, has var- ied, and it has been assigned to both the Cetoth- eriidae and the Balaenopteridae. At least 13 species of Late Miocene and Pliocene age from throughout the world have been assigned to the genus. While it is beyond the scope of the present study to resolve the issues affecting the compli- cated typology and systematics of Plesiocetus and the species that have been assigned to it, it is useful to mention some of the problems that surround it (see also Cabrera 1926; Kellogg 1968:103-104). Both Winge ( 1 9 1 0) and True (1912), who pre- sented a generic diagnosis, considered Plesio- cetus to be a valid balaenopterid genus, but Kel- logg (1968:104) wrote that "Sufficient diagnos- tic criteria for the recognition of Plesiocetus as a valid genus are not presently known." Brandt (1872:407) assigned the genus Plesiocetus to what he regarded as the balaenopterid subfamily Cetotheriinae, the group now recognized as the family Cetotheriidae. Van Beneden (1872:242) fixed the type species as Plesiocetus garopii Van Beneden, 1859, and considered the genus to be related to balaenopterines. Kellogg (1925:51), however, noting that in 1 885 Van Beneden omit- ted P. garopii from a listing of species in the genus Plesiocetus, fixed the type species as Ple- siocetus hupschii Van Beneden, 1859, and dis- cussed it in the context of the family Cetotheriidae. Later (1934:79) Kellogg, re- stricted the genus name Plesiocetus to Plesio- cetus garopii (Van Beneden's type species), and considered it to be a balaenopterine. Comparisons with Plesiocetus cortesii Regardless of the status and systematics of the genus Plesiocetus, one species that has been re- ferred to the genus is of direct relevance to the present study because it helps to establish the polarity of the characters of Parabalaenoptera baulinensis and its taxonomic position. This is the species Plesiocetus cortesii, from Pliocene deposits in Italy, the skull and dentary of which were illustrated by Van Beneden (1875). The skull of Plesiocetus cortesii (see Fig. 8a) resembles that of Parabalaenoptera baulinensis (see Fig. 8b) in its general proportions, but it differs in some details. Compared to Parabalaenoptera baulinensis, P. cortesii (Fig. 8a) has the following more primitive character states: 1 ) The contact between the posterior border of the maxilla and the anterior margin of the su- praorbital process of the frontal is transversely straight rather than curved anteriorly. 2) The postorbital processes of the frontals are larger and project posterolateral ly. 3) The occipital shield is smaller and has more nearly the shape of an equilateral triangle. 4) The zygomatic processes are narrower and straighten 5) The mandibular condyle is smaller and not so spherical. The derived character states of the above five structures (the opposite polarity of the character 136 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 states in Plesiocetus cortesii) are shared by Para- balaenoptera baulinensis and the subfamilies Balaenopterinae and Megapterinae, and indicate that the latter three are sister taxa (Node 3 on the cladogram, Fig. 9). Compared to Parabalaenoptera baulinensis, P. cortesii has nasal bones that are much shorter, terminating at a point posterior to the level of the antorbital notches, and this is a derived character that defines P. cortesii and is listed at Node 2 on the cladogram (Fig. 9). The nasal bones are simi- larly short in Recent Balaenopterinae and Megapterinae, but the nasals of P. cortesii differ by being narrower and not expanded anteriorly. Because of the numerous other primitive charac- ters of P. cortesii, we conclude that short nasal bones evolved convergently in P. cortesii and in both of the derived subfamilies Balaenopterinae and Megapterinae. Plesiomorphies that Plesiocetus cortesii shares with Parabalaenoptera baulinensis include the narrow posterior terminations of the maxillae, the narrow anterior ends of the nasals, and the narrow intertemporal region. Plesiocetus cortesii is a primitive balaenopterid, and in most of its characters is more primitive than Parabalaenop- tera baulinensis. Conclusions 1 . Parabalaenoptera baulinensis, a new genus and species of extinct whale in the family Balaenopteridae, is based on a partial fossil skeleton collected from the sea cliff at Bolinas Point, Point Reyes Peninsula, in Marin County, California. 2. The holotype specimen is from the Late Miocene age Santa Cruz Mudstone, and analysis of diatoms from the rock unit indicates an age of between approximately 6.8 and 6.0 Ma. 3. The holotype skeleton includes the skull, both dentaries, cervical, thoracic, and one lumbar vertebrae, pectoral limb bones, and ribs, making this one of the most complete fossil balaenopterid specimens recorded in the published literature. The animal is estimated to have had a total body length of approximately 10 m in life. 4. Parabalaenoptera baulinensis is a relatively primitive balaenopterid, and it lacks synapomor- phies of the derived balaenopterids of the sub- families Balaenopterinae and Megapterinae. It has a suite of autapomorphies (conspicuously elevated or swollen anterior portions of the pa- rietals on either side of the intertemporal region, long and ascending coronoid crest on the dentary approaching the coronoid process, and posteri- orly-directed apex of the coronoid process), and we classify it in the new subfamily Parabalaenop- terinae. 5. Parabalaenoptera baulinensis shares sym- plesiomorphies with the primitive family Cetoth- eriidae (narrow intertemporal region, larger exposure of frontals on the dorsal surface of intertemporal region, long nasals, long and pos- teriorly tapering ascending processes of the max- illae, and short postglenoid processes of the squamosals), and these shared primitive charac- ters lend support to the theory that the Balaenop- teridae evolved from the Cetotheriidae. 6. Species of the genus Plesiocetus Van Beneden, 1859, pose many problems of classifi- cation and relationships. One species in the ge- nus, Plesiocetus cortesii (Capellini, 1865), is possibly a sister taxon of a less inclusive clade containing Parabalaenoptera baulinensis, the subfamily Balaenopterinae, and the subfamily Megapterinae. 7. Because other, more highly derived taxa of Balaenopteridae are contemporaneous with P. baulinensis, it may be considered to be a relict descendant of a lineage that was ancestral to both the subfamilies Megapterinae and Balaenopteri- nae, and thus to the extant Megaptera and Balaenoptera. The origin of Parabalaenoptera baulinensis must lie among pre-Late Miocene mysticetes that are closer in morphology to primitive members of the family Cetotheriidae. ACKNOWLEDGMENTS The third author regrets to report the deaths of the first two authors, and dedicates this work to the memories of Carl V. Zeigler, who discovered the holotype of Parabalaenoptera baulinensis and was the foundation of the College of Marin whale project, and of Gordon L. Chan, who had a love of the paleontology, geology, and natural history of Marin County and supported the whale project. We thank the many College of Marin students and laboratory assistants who helped excavate and assemble the holotype of Parabalaenoptera baulinensis. We especially acknowledge Ron Melander, who spent many hours working on the ZEIGLER, CHAN, AND BARNES: PARABALAENOPTERA BAULINENSIS 137 holotype and preparing the illustrations, and David Ferguson, the whale project coordinator at College of Marin. We acknowledge consultants William N. Orr of the Geology Department at the University of Oregon; John A. Barron and Ellen J. Moore of the U. S. Geological Survey in Menlo Park, Cali- fornia; and Terrence Gosliner and Peter U. Rodda of the California Academy of Sciences. Marjorie A. Chan, Professor of Geology at the University of Utah, shepherded this manuscript to completion to fulfill her father's wishes. Ac- cordingly, Dr. Chan took care of correspondence related to this publication, helped edit the text, and prepared the illustrations. Finally, we thank the agencies that have funded our work: Marin Community College District, Chancellor's Office of the State of Cali- fornia Community Colleges, The Marin Founda- tion, and a host of private individual contributors. Zeigler and Chan initiated the manuscript and Barnes completed and modified it for publica- tion. Literature Cited Barnes, L. G. 1976. Outline of eastern North Pacific fossil cetacean assemblages. Syst. Zool. 25:321-343. Barnes, L. G., D. P. Domning and C. E. Ray. 1985. Status of studies on fossil marine mammals. Mar. Mamm. Sci. 1(1): 15-53. Barnes, L. G. and S. A. Mcleod. 1984. The fossil record and phyletic relationships of gray whales. Chapter 1 . Pp. 3-32 in The gray whale: Eschrichtius robustus. M. L. Jones, S. L. Swartz, and S. Leather- wood, eds. Academic Press, Orlando, Florida. BENHAM, W. B. 1937. Fossil Cetacea of New Zealand. II — On Lophocephalus, a new genus of zeuglo- dont Cetacea. Trans. Roy. Soc. New Zealand 67:1-7, pis. 1-7. Brandt, J. F. 1872. Bericht uber den bereits vollen- deten druckfertigen Theil seiner Untersuchungen uber die fossilen und subfossilen Cetaceen Europa's. Bull. Acad. Sci. St. Petersburg, Ser. 3, 17:401-408. Cabrera, A. 1926. Cetaceos fosiles del Museo de la Plata. Rev. Mus. La Plata 29:363-4 1 1 . Capellini, G. 1 865. Balenottere fossili del Bolognese. Mem. Accad. Sci. Bologna, Ser. 2, 4:3 1 5-336, 3 pis. Clark, J. C, E. E. Brabb, H. G. Greene and D. C Ross. 1984. Geology of the Point Reyes Peninsula and implications for San Gregorio fault history. Tectonics and sedimentation along the California margin. Soc. Econ. Paleontol. Mineral., Pacific Sect. 38:67-86. Cope, E. D. 1872. On an extinct whale from Califor- nia. Proc. Acad. Natur. Sci. Philadelphia 1 872 (part l):29-30. DEMERE, T. A. 1986. The fossil whale, Balaenoptera davidsonii (Cope, 1 872), with a review of other Neogene species of Balaenoptera (Cetacea: Mys- ticeti). Marine Mammal Science 2(4):277-298. DOMNING, D. P. 1978. Sirenian evolution in the North Pacific Ocean. Univ. Calif. Publ. Geol. Sci. 1 8:i-xi, 1-1 76, pis. 1-18. DU Bus, B. A. L. 1867. Sur quelques mammiferes du Crag d' Anvers. Bull. Acad. Roy. Sci. Belgique, Ser. 2(24):562-577. FORDYCE, R. E. AND L. G. Barnes. 1994. The evolu- tionary history of whales and dolphins. Pp. 4 1 9-455 in 1994 Annual Review of Earth and Planetary Sciences, Volume 22, G. W. Wetherill, ed. Annual Reviews, Inc., Palo Alto, California. Galloway, A. J. 1977. Geology of the Point Reyes Peninsula, Marin County, California. Calif. Div. Mines Geol. Bull. 202:i-v, 1-72, pi. 1 (map). Hanna, G D. and M. E. McLellan. 1924. A new species of whale from the type locality of the Mon- terey Group. Proc. Calif. Acad. Sci., 4th Series, 13(14):23 7-241, pis. 5-9. Hatai, K., S. Hayasaka and K. Masuda. 1963. Some fossil tympanies from the Mizuho period of northern Japan. Saito Ho-on Kai Mus. Natur. Hist. Res. Bull. 32:5-17. Howell, A. B. 1930. Aquatic mammals. Their adap- tations to life in the water. Charles C. Thomas, Springfield, Illinois, and Baltimore, Maryland. K.ELL0GG, A. R. 1922. Description of the skull of Megaptera miocaena, a fossil humpback whale from the Miocene diatomaceous earth of Lompoc, California. Proc. U. S. Natl. Mus. 61(14): 1-1 8, pis. 1-4. [38 PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Volume 50, No. 4 . 1925. Additions to the Tertiary history of pelagic mammals on the Pacific coast of North America. II. Fossil cetotheres from California. Carnegie Inst. Washington Publ. 348(2):35-56. . 1928. The history of whales — their adapta- tion to life in the water. Quart. Rev. Biol. 3(l,2):29-76, 174-208. . 1931. Pelagic mammals from the Temblor Formation of the Kern River region, California. Proc. Calif. Acad. Sci., 4th Series, 1 9( 1 2):2 1 7-397. . 1934. The Patagonian fossil whalebone whale, Cetotheriwn moreni (Lydekker). Carnegie Inst. Washington Publ. 447:63-81, pis. 1-4. — . 1968. Fossil marine mammals from the Mio- cene Calvert formation of Maryland and Virginia. Part 5. Miocene Calvert mysticetes described by Cope. U. S. Nat. Mus. Bull. 247:103-132. Lydekker, R. 1 894. Contributions to a knowledge of the fossil vertebrates of Argentina. 2. Cetacean skulls from Patagonia. An. Mus. La Plata, Sec. Paleontol. (2):1-13, 6 plates. Miller, G. S. 1923. The telescoping of the cetacean skull. Smithsonian Misc. Coll. 75(5): 1-71 , pis. 1-8. -. 1924a. Some hitherto unpublished photo- graphs and measurements of the blue whale. Proc. U.S. Natl. Mus. 66(7): 1-4. -. 1924b. A pollack whale from Florida pre- sented to the National Museum by the Miami Aquarium Association. Proc. U. S. Natl. Mus. 66(9): 1-1 5. Oishi. M. and Y. Hasegawa. 1995. Diversity of Pliocene mysticetes from eastern Japan. The Island Arc [Thematic Issue: Evolution and Biogeography of Fossil Marine Vertebrates in the Pacific Realm, Proceedings of the 29th International Geological Congress, Kyoto, Japan, edited by L. G. Barnes, Y. Hasegawa, and N. Inuzuka] 3(4):436-452 [for 1994]. Omura, H. 1959. Bryde's whale from the coast of Japan. Sci. Rep. Whales Res. Inst. 14:1-33. Repenning,C. A.andR. H.Tedford. 1977.0tarioid seals of the Neogene. U. S. Geol. Surv. Prof. Paper 992:i-vi, 1-93, pis. 1-24. True, F. W. 1 904. The whalebone whales of the west- ern North Atlantic. Smithsonian Contrib. Knowl. 33:1-332. . 1912. The genera of fossil whalebone whales allied to Balaenoptera. Smithsonian Misc. Coll. 59(6): 1-8. Van Beneden, P. J. 1859. Ossements fossiles decou- verts a Saint Nicolas en 1 859. Bull. Acad. Roy. Sci. Belgique, Ser. 2,8(1 1 ): 123-146. 1872. Les baleines fossiles d'Anvers. Bull. Acad. Roy. Sci. Belgique, Ser. 2, 34(7):6-20. . 1875. Le squelette de la baleine fossile du Musee de Milan. Bull. Acad. Roy. Sci. Belgique, Ser. 2,40(12):736-758, figs. 1-2. . 1880. Les mysticetes a courts fanons des sables des environs d'Anvers. Bull. Acad. Roy. Sci. Belgique, Ser. 2, 50(7): 11 -27. 1 882. Description des ossements fossiles des environs d ? Anvers, Troisieme Partie, Cetaces. Gen- res: Megaptera, Balaenoptera, Burtinopsis & Erpe- tocetus. Ann. Mus. Roy. d'Hist. Nat. Belgique, Ser. Paleontologique 7:1-90, pis. 40-109. . 1 885. Description des ossements fossiles des environs d'Anvers, Quatrieme Partie, Cetaces. Genre: Plesiocetus. Ann. Mus. Roy. d'Hist. Nat. Belgique, Ser. Paleontologique 9:1^40, pis. 1-30. Wilson, L. E. 1937. New species of dolphin from Point Reyes, California. Abstract, Proc. Geol. Soc. Amer. 1936:387. WlNGE, H. 1910. Om Plesiocetus og Sqvalodon [sic] fra Danmark. Vidensk. Meddel. Naturhist. Foren., Copenhagen, 7(1): 1-38, pis. 1-2. CALIFORNIA ACADEMY OF SCIENCES, 1997 Golden Gate Park San Francisco, California 941 18