Extraordinarily thick-boned fish linked to thearidification of the Qaidam Basin (northernTibetan Plateau)Meemann Chang*†, Xiaoming Wang‡, Huanzhang Liu§, Desui Miao¶, Quanhong Zhao�, Guoxuan Wu�, Juan Liu*,Qiang Li*, Zhencheng Sun**, and Ning Wang*

*Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, P.O. Box 643, Beijing 100044, China; ‡Department ofVertebrate Paleontology, Natural History Museum of Los Angeles County, Los Angeles, CA 90007; §Institute of Hydrobiology, Chinese Academy of Sciences,Wuhan 430072, China; ¶Natural History Museum and Biodiversity Research Center, University of Kansas, Lawrence, KS 66045; �Laboratory of MarineGeology, Tongji University, Shanghai 200092, China; and **Department of Earth Sciences, China University of Petroleum–Beijing, Beijing 102249, China

Communicated by Peter R. Crane, University of Chicago, Chicago, IL, June 27, 2008 (received for review October 10, 2007)

Scattered with numerous salt lakes and �2,700–3,200 m above sealevel, the giant Qaidam inland basin on the northern TibetanPlateau has experienced continuing aridification since the begin-ning of the Late Cenozoic as a result of the India–Asia plate collisionand associated uplift of the Tibetan Plateau. Previous evidence ofaridification comes mainly from evaporite deposits and salinity-tolerant invertebrate fossils. Vertebrate fossils were rare untilrecent discoveries of abundant fish. Here, we report an unusualcyprinid fish, Hsianwenia wui, gen. et sp. nov., from Pliocene lakedeposits of the Qaidam Basin, characterized by an extraordinarilythick skeleton that occupied almost the entire body. Such enor-mous skeletal thickening, apparently leaving little room for mus-cles, is unknown among extant fish. However, an almost identicalcondition occurs in the much smaller cyprinodontid Aphaniuscrassicaudus (Cyprinodonyiformes), collected from evaporites ex-posed along the northern margins of the Mediterranean Seaduring the Messinian desiccation period. H. wui and A. crassicaudusboth occur in similar deposits rich in carbonates (CaCO3) andsulfates (CaSO4), indicating that both were adapted to the extremeconditions resulting from the aridification in the two areas. Theoverall skeletal thickening was most likely formed through depo-sition of the oversaturated calcium and was apparently a normalfeature of the biology and growth of these fish.

Increasing evidence ascribes a profound climatic shift, such asAsian monsoon systems and widespread aridification, to theuplift of the Tibetan Plateau caused by the India–Asia platecollision during the Late Cenozoic. The cyprinid fish reportedhere (Fig. 1) are from the lacustrine deposits of the PlioceneShizigou Formation in the northeastern wing of Yahu Anticline,Qaidam Basin, northern Tibetan Plateau (1) [Fig. 2A and formore details see supporting information (SI) Fig. S1], collectedfrom 20 localities in a 220-m-thick sequence of siltstones, finesandstones, and marls under a marker bed, Horizon K (2) (Fig.2B). The deposits are rich in carbonates (CaCO3) and sulfates(CaSO4). The unusually thick-boned fish represents an adaptivemode unknown in any extant fish and extremely rare in fossils.Based on this fossil fish, our morphological description andcladistic analysis shed light on schizothoracin phylogeny. We alsosuggest that the overall skeletal thickening was formed throughdeposition of the oversaturated calcium and reflects the fish’sdramatic physiological adjustment to severe environmental dis-tress. The fish thus witnessed the process of aridification andprovided a convincing link between environmental changes onthe Tibetan Plateau and biological responses of its inhabitants.

Systematic PaleontologyOstariophysi Sagemehl, 1885. Cypriniformes Bleeker, 1860; Cyp-rinidae Bonaparte, 1837; Schizothoracinae Berg, 1912; Hsian-wenia wui, gen. et sp. nov.

Holotype. Nearly complete skeleton, only with posterior part ofthe caudal fin and a small anterodorsal portion of the bodymissing, IVPP V 15244, from locality CD0649, Fig. 3 A–C.Included material. V 15012, nearly complete skeleton with the tail,a short portion between the body and head, and the snoutmissing (Figs. 1 A and B and 3D) from locality CD0507; V15245.1 and V 15245.2, both are middle part of the fish skeleton,from locality CD0649; V 15306.1–37, detached bones of headand body, V 15307.1–50, detached pharyngeal teeth, fromlocalities CD0506 and CD0507.Etymology. The genus and species name is dedicated to the lateProf. Hsianwen Wu, who made great contributions to the studyof the Chinese cyprinids.Horizon and localities. Localities CD0504–0507, CD0641–0653,CD0663–0666, upper part of the Shizigou Formation, YahuAnticline, Qaidam Basin, Qinghai Province, China, N37°44�38�–46�22�, E93°36�55�–38�38�; Pliocene.Diagnosis. Schizothoracine with a unique combination of charac-ters shared with barbines and primitive schizothoracines on onehand and specialized schizothoracines on the other: body elon-gated, nearly cylindrical; head relatively long; orbit small, situ-ated anteriorly; anterior margin of lower jaw shovel-like; pha-ryngeal bone stout, triangle-shaped, with three rows ofcylindrical teeth, new teeth with rounded top, worn teeth withtruncated, f lat grinding surface; distal end of transverse processof 4th vertebra blunt and slightly expanded; fork in pelvic boneshallow, both branches comparatively broad; dorsal fin withthree unbranched and seven branched rays, the longest un-branched ray with robust serrations; proximal portion of epuralexpanded.Description and remarks. The body (Figs. 1 A and B and 3A) iselongated, more or less rounded anteriorly and slightly com-pressed posteriorly. The standard length in the holotype V 15244is 5.4 times that of the body depth (78 mm). The head is relativelylong (110 mm), 1.7 times its depth (65 mm), and nearly a quarterof the standard length. All specimens, including the disarticu-lated bones, are comparatively large. The holotype V 15244 hasa standard length (length from the tip of snout to base of caudalfin) of 425 mm, whereas the estimated standard length of V15012 is �445 mm. Body length with the caudal fin is estimatedas 450–500 mm. Judged from the preserved parts of the body, V15245.1 and V 15245.2 must have been even larger. The growth

Author contributions: M.C. and X.W. designed research; X.W. led the expeditions; M.C.,H.L., J.L., Q.L., and N.W. performed research; M.C., H.L., Q.Z., G.W., and Z.S. analyzed data;and M.C. and D.M. wrote the paper.

The authors declare no conflict of interest.

†To whom correspondence should be addressed. E-mail: cmeemann@yahoo.com.

This article contains supporting information online at www.pnas.org/cgi/content/full/0805982105/DCSupplemental.

© 2008 by The National Academy of Sciences of the USA

13246–13251 � PNAS � September 9, 2008 � vol. 105 � no. 36 www.pnas.org�cgi�doi�10.1073�pnas.0805982105

Dow

nloa

ded

by g

uest

on

June

12,

202

1

http://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/content/full/0805982105/DCSupplementalhttp://www.pnas.org/cgi/content/full/0805982105/DCSupplemental

rate of recent schizothoracines is very slow. Those with a bodylength 400 mm typically have an age of at least 9–10 years (3).Thus, most of these schizothoracines were probably 10 (or even15) years old.

The structure of the skull roof, the opercular system, and thepalate is similar to that of other cyprinids. Judged from theposition of the infraorbital bones, the orbit is quite small andsituated rather anteriorly. The postorbital distance (distancebetween posterior margin of opercle and posterior margin oforbit) is long, more than four times that of the antorbital distance(distance between the anterior margin of the orbit and tip ofsnout) (Fig. 3C). Among recent schizothoracines, those living inbroad rivers or lakes often use the caves beneath the sod-heldbank edges as refuge during their winter hibernation. In summer,they also hide in the caves during the day, and only come out atdusk to search for food (3). It is therefore likely that their eyesbecame reduced and their sight weakened, like some cave-livingbarbines today.

The mouth is subterminal, with the upper jaw protrudingbeyond the lower (Fig. 3C). A shovel-like anterior margin of thedentary was apparently covered with a horny sheath when thefish was alive and is related to the habit of feeding on algaeattached to the bottom. As in most cyprinids, the anteriorprocess of premaxilla and coronoid process of the dentary arewell developed. The quadrate has a very shallow notch forarticulation of symplectic and is in contact with the metaptery-

goid, leaving no opening between them (Fig. 3C). The ectoptery-goid is deep and narrow. The pharyngeal bone is stout andtriangle-shaped (Fig. 3E). There are three rows of teeth on eachbone, with a pattern of 2.3.5–5.3.2. The teeth are cylindrical inshape; new teeth have rounded top, whereas the worn teeth aretruncated with flat grinding surface (Fig. 3F).

The total number of vertebrae (�46) is greater than inprimitive barbines. The dorsal fin has three unbranched andseven branched rays. The posterior edge of the longest un-branched ray is equipped with gross serrations (Figs. 1 A and 3A).The anal fin has three unbranched and five branched rays. Theposterior edge of the longest unbranched ray is smooth. Thepelvic fin is situated more posteriorly than in primitive barbines.Its origin is under the 5th dorsal fin ray. The fork in the anteriorpart of the pelvic bone is less than half of the main part of thebone, and both branches are broad. The structure of the caudalskeleton (Fig. 3B) resembles that of most cyprinids. The usuallyslender, singular epural of cyprinids is widely expanded in thisfish (Fig. 3B).

The most striking aspect of the fish specimens from the YahuAnticline is that almost all bones from the skull, paired fingirdles, vertebrae with their arches and spines, pterygiophores,and even all intermuscular bones and gill rakers (Figs. 1 A andB and 3 A–I), are extremely thickened. On specimen V 15012, theribs and epineurals in anterior portion of the body grow so closeto each other that the vertebrae in this region are completely

Fig. 1. H. wui gen. et sp. nov. (A) Body without head and tail (IVPP V 15012) in right view. (B) The same specimen in left view, revealed by CTScan, flippedhorizontally. (C) Gymnocypris przewalskii Kessler, radiograph of body in right view. (D) box area in C in higher magnification. (E) A. crassicaudus (Agassiz), middleportion of body in left view, after Gaudant (25). (Scale bars: 30 mm in A–C); 3 mm in E.) an, anal fin; ds, dorsal fin spine; epn, epineurals; epp, epipleurals; hs,haemal arches and spines; ns, neural arches and spines; ptr, pterygiophores; ri, ribs; ver, vertebrae.

Chang et al. PNAS � September 9, 2008 � vol. 105 � no. 36 � 13247

EVO

LUTI

ON

ENV

IRO

NM

ENTA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

June

12,

202

1

covered by them and cannot be seen from the lateral side (Fig.1 A and B). The parapophysis of the 4th vertebra is very robust(Fig. 3G), whereas parapophyses of other vertebrae are also verythick and often preserved in disarticulated state. This makesmany isolated bones look odd and difficult to recognize at firstglance (Fig. 3 G–I). The bones in V 15244 (the smallest amongthe available specimens) are obviously thinner than those inthree larger specimens V 15012, V 15245.1, and V 15245.2(compare Fig. 1 A and B, Fig. 3A, and Fig. S2). This indicates thatthe thickness of the bone appears to increase with age.

Results and DiscussionPhylogenetic Position of H. wui. H. wui resembles barbines andprimitive schizothoracines (Cyprinidae) in many respects and, atthe same time, shares a few characters with some specializedschizothoracines. The schizothoracines [Schizothoracin lin-eage � Oreininae, Cyprininae (4), Tribe Oreinini, Cyprininae(5), Cyprinidae] are endemic to the aquatic systems of theTibetan Plateau and surrounding areas (including southwesternChina, Kazakhstan, Kyrgyzstan, Tajikistan, Afghanistan, Paki-stan, India, Nepal, and Bhutan) (5–8). The Cyprinidae are thelargest and most widespread group of freshwater fish and are ofconsiderable significance in fisheries. They comprise 2,420 spe-cies (9) distributed on all continents except Australia, SouthAmerica, and Antarctica. Although the group has been studiedfor nearly two centuries, views about its subdivision still diverge.Schizothoracines are regarded as either a subfamily of theCyprinidae (6, 10–12) or a monophyletic lineage within thesubfamily Cyprininae (sensu lato, including the Barbinae) (4, 5)or Barbinae (sensu lato) (9, 13). The origin and diversification ofschizothoracines is thought to be closely related to the uplift ofthe Tibetan Plateau. The recent schizothoracines have beendivided into three categories: primitive, specialized, and highlyspecialized, based on the extent of modifications of their scales,barbells, and rows of pharyngeal teeth and their distribution atthree successive altitudes at which water temperature and pre-cipitation decrease, and solar radiation and evaporation increase(8). Phylogenetic studies of the Schizothoracinae using cladisticmethod have arrived at similar conclusions (14, 15).

Fossil schizothoracines are extremely rare. Relatively com-

plete specimens from the Neogene Dingqing Formation ofLunbola Basin, northern Tibet, that are similar to schizotho-racines were described as a new genus and species Plesioschizo-thorax macrocephalus and referred to the Cyprinidae (16).Incomplete pharyngeal bones and disarticulated pharyngealteeth have also been reported from early Oligocene to earlyPliocene deposits of East Kazakhstan, the mountainous regionof Altai, and Mongolia (17). Information about these fossilsremains sketchy. They indicate the presence of possibly relatedfish in these areas during earlier time intervals. The Yahuschizothoracines are the best preserved to date. The localitieswhere they occur and the Eboliang locality, where unidentifieddisarticulated fish bones were found, are just outside the distri-bution of the extant schizothoracines.

To examine the phylogenetic position of H. wui, we compiled adataset with 12 taxa and 32 characters (see Table S1). The 12 taxainclude 7 extant schizothoracines and 2 fossil forms, plus 3 cyprininand barbinin genera serving as outgroups. Characters were takenmainly from the literature (14, 15), augmented with our ownobservations of both extant and fossil forms. The phylogeneticanalysis using PAUP Version 4.0b 10 (18) resolves Plesioschizotho-rax macrocephalus and H. wui among the early branching schizo-thoracines, in a position slightly more derived than Schizothorax.The remaining schizothoracines retain positions indicated by pre-vious authors (Fig. 4; see also SI Text). In this case, the two fossilforms, H. wui and Plesioschizothorax macrocephalus, are not themost ancient members of the group; they are somewhat in betweenthe primitive members of the group, e.g., the extant genus Schizo-thorax and more derived members. The history of the group wouldgo back to a time earlier than the Neogene. Our analysis shows thatschizothoracines form a monophyletic group within the Barbinae(sensu lato) (11, 13) indicating that the Barbinae sensu stricto (11, 12,19) is paraphyletic.

Interpretations of the Skeletal Thickening. Swollen or partially swol-len bones are not uncommon in recent tropical and subtropicalmarine teleosts and are observed in taxonomically diverse groups(20). The phenomenon is referred to as hyperostosis (21), that is,the usually thin bones or parts of bones expand into rounded, bulkystructures. The swollen parts only occur in a few bones in the samefish (see Fig. S3), and although their location varies from species tospecies, they always occur on the same bones in a given species (20,21). Similar structures have also been reported in fossil fish fromEurope, North America, and Africa, known as ‘‘Tilly Bones’’(21–23), and are presumed to have been caused by disease or aging(21) or by abnormal hydrochemical conditions, such as high contentof fluorine in the water (24).

The overall thickening of the skeleton in the Yahu schizotho-racine H. wui is not comparable with that in the recent marinefishes nor is it comparable with the ‘‘Tilly bones’’ of other fossils.A similarly thickened skeleton has been observed only in thecyprinodont fossil Aphanius crassicaudus (Fig. 1E) recoveredfrom the late Miocene Messinian evaporites exposed along thenorthern coast of the Mediterranean (southeastern Spain, Sicily,and Crete) (25–27). A. crassicaudus was first described as a fossilspecies of the extant genus Lebias and later reassigned to thefossil genus Pachylebias (25). It is now referred to as the genusAphanius (27), an extant multispecies genus that has a circum-Mediterranean distribution and is adapted to varying degrees ofsalinities (25). The authors referred to the bone hypertrophy aspachyostosis.

Although the two known pachyostotic species, H. wui and A.crassicaudus, are of different age (Pliocene and late Miocenerespectively) and are from different environments (lacustrineversus marine or lagoonal), they share many features in common.Both show an overall thickening of skeleton, which is not knownin any other fish. This thickening apparently had little adverseinfluence on the life of the fish, and all individuals seem to have

Fig. 2. Localities and section. (A) The studied area, Yahu and Eboliang inQaidam Basin. (B) Yahu section, numbers to the left of the column, thicknessof the section; numbers to the right of the column, the numbers of therelevant localities. CD, Chaidam (Qaidam), in the numbering system used byWang et al. (2). (Scale bar: 1,000 km.)

13248 � www.pnas.org�cgi�doi�10.1073�pnas.0805982105 Chang et al.

Dow

nloa

ded

by g

uest

on

June

12,

202

1

http://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=SF2http://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=STXThttp://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=SF3

lived a relatively long, and normal, life (growing to more thanhalf a meter long in the case of H. wui) even with increasinglythickened skeleton. In both cases, thickening of bones seems toappear when the fish reached a certain size (or age) and increaseswith increasing size (or age). The timing of bone thickening inA. crassicaudus varies from locality to locality (26). Although wedo not have specimens of H. wui smaller than 445 mm in length(V 15244), we did notice that the larger the fish the morepronounced the degree of the bone thickening. Hence, it isconceivable that the bones in H. wui might not have beenthickened at all when the fish was below a certain size.

The fish fauna associated with both species at most localitiesis dominated by one species, suggesting unfavorable conditionsfor a more diversified fish fauna to survive. Perhaps restrictedmovement resulting from the thickened bones may not have beendisadvantageous because of the lack of predators in the envi-ronment. It may also be significant that the recent relatives ofboth H. wui and A. crassicaudus, various species of Aphanius andsome schizothoracines (e.g., Gymnocypris przewalskii), live invaried salinities (euryhaline), from water with high salinity tobrackish or even completely fresh water.

The geological occurrence of H. wui and A. crassicaudus is alsosimilar. Both were collected from deposits rich in marl (CaCo3)and gypsum (CaSo4), indicating similar hydrochemical condi-tions with a high calcium content. A. crassicaudus occurs inmarl–gypsum along the coastal region of the Mediterraneandeposited during the Messinian Salinity Crisis (28), where theless soluble carbonates and sulfates precipitated first, and the

more soluble sodium chloride and other minerals deposited laterin more central part of the basin. In the deposits of the YahuAnticline, carbonate ooliths and crystals of gypsum are oftenseen in the rock samples, also documenting a high concentrationof calcium. Geological evidence also suggests that both taxasurvived for a significant period. The �220-m-thick fish-bearingsediments from the Yahu Anticline indicate that Hsianweniaexisted for a period of up to 200,000 years (2) (Fig. 2B) beforea phase of general desiccation, although the Mediterranean areawas later flooded by the Atlantic marine water, whereas theQaidam Basin went on to develop into an arid badland.

The palynological assemblage recovered from the fish-bearing deposits from the Yahu Anticline, mainly includesPicea, Abies, Pinus, Artemisia, Chenopodiaceae, and Gra-mineae, and documents a vegetation closely similar to that ofarid area of northwestern China today. Pediastrum, an algausually widespread in shallow freshwaters, is also frequentlyobserved. In the deposits of Eboliang, there is a high percent-age of Typha (82.9%), a plant often seen along the margin oflakes and rivers in northern China (see SI Text and Table S2).Ostracods from the fish-bearing strata are Ilyocypris bradyi,Eucypris sp., Leucocythere mirabilis, and Microlimnocytherereticulata. The first is a freshwater species, whereas the re-maining species are typical brackish forms. Ilyocypris bradyi,certain species of Eucypris, and Leucocythere mirabilis inhabitthe present day Qinghai Lake just to the east to our fossil fishlocalities. The present Lake Qinghai has a salinity of 14‰ andis rich in magnesium sulfate. The content of sodium chloride

Fig. 3. H. wui gen. et sp. nov. (A) Holotype (IVPP V 15244) in right view. (B) Caudal skeleton of holotype in right view. (C) Head of holotype in right view. (D) Ossifiedgill rakers (IVPP V 15012). (E) Pharyngeal bone with teeth in ventromedial view (IVPP 15245.1). (F) Four pharyngeal teeth, top view, showing grinding surface (left) andposterior view (right) (from left to right: IVPP V 15307.2, 7, 4, 14). (G) Left parapophysis of 4th vertebra in ventromedial (left) and dorsolateral (right) view (IVPP V 15306.33). (H) Two disarticulated epineurals (left, IVPP V 15306.20; right, IVPP V 15306.18). (I) Disarticulated epipleural (IVPP V 15306.19). (Scale bars: 20 mm in A and C; 10mm in B, E, G, H, and I; 5 mm in D and F.) an, anal fin; d, dentary; ds, dorsal fin spine; ect, ectopterygoid; ent, entopterygoid; ep, epural; epn, epineurals; epp, epipleurals;fr, frontal; hs, haemal arches and spines; hy, hyomandibular; iop, interopercle; io, infraorbitals; mpt, metapterygoid; mx, maxilla; ns, neural arches and spines; op,opercle; pa, parietal; pop, preopercle; pmx, premaxilla; ptr, pterygiophores; q, quadrate; ri, ribs; un, uroneural; ver, vertebrae.

Chang et al. PNAS � September 9, 2008 � vol. 105 � no. 36 � 13249

EVO

LUTI

ON

ENV

IRO

NM

ENTA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

June

12,

202

1

http://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=STXThttp://www.pnas.org/content/vol0/issue2008/images/data/0805982105/DCSupplemental/SD1.xls

is the highest (66%), whereas MgSO4 comes second (20%).Despite this relatively high salinity, but perhaps because of themoderate calcium content, the single schizothoracin speciesnow living in the lake, G. przewalskii, has normal skeletonswithout any thickened bones (Fig. 1 C and D). Similarly,radiographs of a dozen specimens of the extant species A.iberus showed no abnormal bone thickening, although thespecimens were collected from the drained surface of a saltlake Sebkha-el-Melah in the Sahara and preserved in salt(NaCl) (26). Thus, high salinity alone without a high concen-tration of calcium may not cause bone thickening.

Several interpretations have been presented for the skeletalthickening in A. crassicaudus (25, 26), including aging, regu-lation of equilibrium related to unstable salinity, high contentof particular chemical elements in the water, or a combinationof a number of these factors. Based on comparison with H. wui,high content of particular chemical elements in the waterseems to be the most likely explanation. And because ofgrowth of fish appears to have been unaffected, the chemicalelements should at least meet two prerequisites: (i) they areharmless to the fish and (ii) they can form a component of thebone.

A high concentration of calcium in the water of the two areaswhere H. wui and A. crassicaudus lived was most probably thecrucial chemical element causing the bone thickening. In bothcases, concentrations of carbonate (CaCO3) and sulfate(CaSO4) were apparently high. Calcium is the main componentof the bone, and it is not harmful, or at least not lethal, to fishin general. This is supported by the thickened cranial bonesdescribed for the cichlid Tilapia guinasana, endemic to Lake

Guinas in northern Namibia. In this case, hyperostosis wasattributed to the environmental induction rather than togenetic reasons. The lake water has very high calcium carbon-ate content (calcium hardness, as CaCO3, 185 ppm). Obser-vations on four specimens of T. guinasana originally derivedfrom wild-caught parents but spawned and raised in laboratorytap water (calcium hardness, as CaCO3, �45 ppm) showthat their roofing bones are thinner than those of wild con-specifics (29).

Under natural condition, when the concentration of calciumis high, the amount of other minerals also tends to be higher thanusual. Therefore, only euryhaline fishes are able to survive fora relatively extended period. Both H. wui and A. crassicauduswere the only species surviving high salinity, but they wereprobably also able to deposit the oversaturated calcium in theirbones and thus form the overall thickened skeleton. The phys-iological mechanism by which the overdeposition of the calciumin the bones occurs remains to be established.

In the Upper Youshashan Formation (stratigraphically un-derlying the Shizigou Formation) in the Eboliang area, whereTypha constitutes 82.9% of the total pollen count, indicatinga freshwater or low-salinity environment (Table S2), thedisarticulated fish bones are normal, not thickened at all. Inthe area of the Yahu Anticline, overlying the fish-bearingShizigou Formation, is the Qigequan Formation, in which noremains of fish have been found. Instead, layers of puregypsum are clearly visible. The transition from the UpperYoushashan Formation with normal-boned fossil fish, to theShizigou Formation with thick-boned fish and then to thebarren Qigequan Formation, appears to show progressivearidification in the area during the Late Cenozoic. Thisunusually thick-boned fish was apparently an eyewitness to, orperhaps a victim of, the drying of Qaidam Basin.

Materials and MethodsMost of the fish remains reported here are from the lacustrine deposits ofthe Pliocene Shizigou Formation in the northeastern wing of Yahu Anti-cline, Qaidam Basin. Most of the material consists of disarticulated bonesand pharyngeal teeth, but two nearly complete and two partially preservedskeletons were also recovered (localities CD0507 and CD0649). A fewdisarticulated, nonthickened fish bones were also recovered from thelatest Miocene to early Pliocene Upper Youshashan Formation exposed inthe axis of Eboliang Anticline in the centrowestern area of the Basin(Fig. 2B; for more details see Fig. S1). Fish specimens were preparedmechanically.

ACKNOWLEDGMENTS. We thank J. Gaudant for sharing personal observa-tions, references, and interesting insights; W. F. Smith-Vaniz for allowing us touse figure 1B in Smith-Vaniz et al. (1995); F. Meunier and E. Sychevskaya fordiscussions; E. M. Friis and three other anonymous reviewers for their criticalreview of the manuscript and helpful suggestions; Z. D. Qiu for providingprecious fossils; F. C. Zhang and W. Gao for help in preparation of illustrations;and Z. Wang for preparing fossils. This work was supported by the NationalNatural Science Foundation of China (40432003) and the U.S. National ScienceFoundation (CToL EF0431326 and EAR-0446699).

1. Zhang PX (1987) Salt lakes of Qaidam Basin (Science Press, Beijing), pp 1–242 (inChinese).

2. Wang XM, et al. (2007) Vertebrate paleontology, biostratigraphy, geochronology, andpaleoenvironment of Qaidam Basin in northern Tibetan Plateau. Palaeogeogr Palaeo-climatol Palaeoecol 254:363–385.

3. Tsao WS, Wu HW (1962) An investigation of the fish biology and fishery problems inGanze-Apa region of western Szecguan Province. Acta Hydrobiol Sin 2:79–110 (inChinese).

4. Howes G (1991) Systematics and biogeography: An overview. Cyprinid Fishes, System-atics, Biology and Exploitation. Fish and Fisheries Series, eds Winfield IJ, Nelson JS(Chapman & Hall, New York), Vol 3, pp 1–33.

5. Kullander SO, Fang F, Delling B, Åhlander E (1999) The fishes of the Kashmir Valley.River Jhelum, Kashmir Valley. Impacts on the Aquatic Environment, ed Nyman L(Swedmar, Göteborg, Sweden), pp 99–168.

6. Berg LS (1912) Fauna of Russia and Adjacent Countries (Imperial Academy of Sciences,St. Petersburg, Russia), Vol 3, pp 369–704.

7. Hora SL (1953) Fish distribution and Central Asian orography. Curr Sci 22:93–97.8. Cao WX, Chen YY, Wu YF, Zhu SQ (1981) Origin and evolution of Schizothoracine fishes

in relation to the upheaval of the Qinghai-Xizang Plateau. Studies on the Period,Amplitude and Type of Uplift of the Qinghai-Xizang Plateau, ed The comprehensivescientific expedition to the Qinghai-Xizang Plateau, Chinese Academy of Sciences(Science Press, Beijing), pp 118–130 (in Chinese).

9. Nelson JS (2006) Fishes of the World, 4th Ed (Wiley, New York), pp 139–148.10. Wu XW (1964) The Cyprinid Fishes of China (Shanghai Science and Technology Press,

Shanghai), Vol 1, pp 1–228 (in Chinese).11. Bãnãrescu P (1992) Distribution and dispersal of freshwater animals in North America

and Eurasia. Zoogeography of Freshwaters (Aula, Wiesbaden, Germany), Vol 2, pp512–1091.

12. Yue PQ, et al. (2000) Fauna Sinica, Osteichthys, Cypriniformes (Science Press, Beijing),Vol 3, pp 273–390 (in Chinese).

13. Cunha C, Mesquita N, Dowling TE, Gilles A, Coelho MM (2002) Phylogenetic relationshipsof Eurasian and American cyprinids using cytochrome b sequences. J Fish Biol 61:929–944.

Fig. 4. Cladogram showing the phylogenetic position of H. wui gen. et sp.nov. among schizothoracines and related taxa.

13250 � www.pnas.org�cgi�doi�10.1073�pnas.0805982105 Chang et al.

Dow

nloa

ded

by g

uest

on

June

12,

202

1

http://www.pnas.org/content/vol0/issue2008/images/data/0805982105/DCSupplemental/SD1.xlshttp://www.pnas.org/cgi/data/0805982105/DCSupplemental/Supplemental_PDF#nameddest=SF1

14. Wu YF (1984) Systematic studies on the cyprinid fishes of the subfamily Schizothoraci-nae from China. Acta Biol Plateau Sin 3:119–140 (in Chinese).

15. Chen YF (1998) Phylogenetic and distributional patterns of subfamily Schizothoracinae(Pisces: Cyprinidae) I. The phylogenetic patterns. Acta Zootaxon Sin 23 supple-ment:17–25 (in Chinese).

16. Wu YF, Chen YY (1980) Fossil cyprinid from the late Tertiary of North Xizang, China.Vertebr Pal Asiat 18:15–20 (Chinese).

17. Sytchevskaya EK (1989) Neogene freshwater fish Fauna of Mongolia. Transact JointSoviet–Mongolian Palaeotol Expedition 39:77–82.

18. Swofford DL (2000) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Meth-ods) Version 4.0b (Sinauer Associates, Sunderland, MA).

19. Wu XW (1977) The Cyprinid Fishes of China (Shanghai People’s Press, Shanghai), Vol2, pp 229–383 (in Chinese).

20. Smith-Vaniz WF, Kaufman LS, Glowacki J (1995) Species-specific patterns of hyperos-tosis in marine teleost fishes. Mar Biol 121:573–580.

21. Konnerth A (1966) Tilly bones. Oceanus 12:6–9.22. Hewitt RA (1983) Teleost hyperostose; a case of Miocene problematica from Tunisia.

Tertiary Res 5:63–70.

23. Schlüter T, Kohrling R, Mehl J (1992) Hyperostotic fish bones (‘‘Tilly Bones’’) from presumablyPliocenephosphoritesof theLakeManyaraarea,northernTanzania.PaläontolZ66:129–136.

24. Schlüter T, Kohring R (2002) Paleopathological fish bones from phosphoritesof the Lake Manyara area, northern Tanzania—fossil evidence of a physiologicalresponse to survival in an extreme biocenosis. Environ Geochem Health24:131–140.

25. Gaudant J (1979) ‘‘Pachylebias’’ crassicaudus (Agassiz) (Poisson Téléostéen, cyprin-odontiforme), un constituant majeur de l’ichthyofaune du Messinien continental dubassin méditerranéen. Geobios 12:47–73.

26. Meunier FJ, Gaudant J (1987) Sur un case de pachyostose chez un poisson du Miocéneterminal du bassin méditerranéen, Aphanius crassicaudus (Agassiz), (Teleostei, Cypri-nodontidae). C R Acad Sci Paris (Sér 2) 305:925–928.

27. Sorbini L, Tirapelle R (1979) Messinian fossil fish of the Mediterranean. PalaeogeogrPalaeoclimatol Palaeoecol 29:143–154.

28. Hsü KJ, Ryan WBF, Cita M-B (1973) Late Miocene desiccation of the Mediterranean.Nature 242:240–244.

29. Greenwood PH (1992) A redescription of the uniquely polychromatic African cichlidfish, Tilapia guinasana Trewavas. Bull Br Mus Nat Hist (Zool) 58:21–36.

Chang et al. PNAS � September 9, 2008 � vol. 105 � no. 36 � 13251

EVO

LUTI

ON

ENV

IRO

NM

ENTA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

June

12,

202

1