D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

PROCEEDINGS O F ---------T H E R O Y A L S O C I E T Y

rspb.royalsocietypublishing.org

Research 3 ®CrossMark

Cite this article: P y en so n ND et al. 2 0 1 4 R e p e a te d m a s s s tra n d in g s o f M io c e n e m a rin e m a m m a ls fro m A ta c a m a R egion o f C hile p o in t t o s u d d e n d e a th a t s e a . Proc. R. Soc. B 281:

2 0 1 3 3 3 1 6 . http://dx.doi.O rg/10.1098/rsp b .2013 .3316

R eceived: 19 D e c e m b e r 2 0 1 3 A cc ep ted : 2 3 J a n u a r y 2 0 1 4

Subject Areas:

e c o lo g y , p a la e o n to lo g y

Keywords:

ta p h o n o m y , s tra n d in g s , fossil reco rd , h a rm fu l a lg a l b lo o m s

Author for correspondence:

N icholas D. P y en so n e -m a il : p y e n so n n @ si.e d u

Repeated mass strandings of Miocene marine mammals from Atacama Region of Chile point to sudden death at seaN icholas D. P y en so n 1-3-4, Carolina S. G utste in1-5, Jam es F. Parham 6, Jacobus P. Le Roux7, Catalina Carreño Chavarria7, H olly L ittle1, A dam M éta llo 8, V incent Rossi8, Ana M. V alenzuela-T oro5-9, Jorge V elez-Ju arb e1-10,Cara M. S an te lli2, David Rubilar R ogers5-11, M ario A. C ozzuol12 an d M ario E. Suárez5-9

d e p a r t m e n t o f P a leobio logy , N ational M useum o f N atural History, a n d d e p a r t m e n t o f M ineral Sciences,N ational M useum o f N atural History, S m ith so n ian In s titu tio n , PO Box 3 7 01 2 , W a sh in g to n , DC 20 01 3 , USAd e p a r t m e n t o f M am m alo g y , an d d e p a r t m e n t o f P aleo n to lo gy , Burke M useum o f N atural History a n d C ulture,S e a ttle , WA 9 8 19 5 , USA5 Red P a leo n to ló g ica , L aborato rio d e O n to g en ia y F ilogen ia, D e p a rta m e n to d e Biología, F acultad d e Ciencias, U niversidad d e Chile, Las P a lm eras, S an tiag o 34 25 , Chile6Jo h n D. Cooper A rchaeological a n d P aleon to log ical C en ter, D e p a r tm e n t o f G eological Sciences, California S ta te University, F u llerton , CA 92 83 4 , USAd e p a r t a m e n t o d e G eo log ía, F acultad d e C iencias Físicas y M a tem á ticas a n d A ndean G eo th erm al C en ter o f Excellence, U niversidad d e Chile, P laza Ercilla 803 , S an tiag o , Chiled ig i t iz a t io n P rogram Office 3D Lab, Office o f th e C hief In fo rm ation Officer, S m ith so n ian In s titu tion , L andover, MD 20 78 5 , USA9 L aborato rio d e Ecofisiología, D e p a rta m e n to d e Ecología, Facultad d e C iencias, U niversidad d e Chile, Las P a lm eras, S an tiag o 34 25 , Chile1 0 Florida M useum o f N atural History, University o f F lorida, G ainesville , FL 3 2 61 1 , USA11Área P a leo n to lo g ía , M useo Nacional d e Historia N atural, Casilla 78 7 , S an tiag o , Chile ’d e p a r t a m e n t o d e Z oología, In s ti tu to d e Ciencias Biológicas, U n iversidade F ederal d e M inas Gerais,Belo H orizon te , M inas G erais, Brazil

M arine m am m al m ass strand ings have occurred for m illions of years, b u t their orig ins defy singu lar explanations. Beyond h u m a n causes, m ass strand ings have been a ttrib u ted to h erd ing behav iour, large-scale oceanographic fronts a n d harm fu l algal b loom s (HABs). Because algal toxins cause o rgan failure in m arine m am m als, HABs are th e m ost com m on m ass s trand ing agen t w ith b ro ad geographical a n d w id esp read taxonom ic im pact. T oxin-m ediated m ortalities in m arine food w ebs have the po ten tia l to occur over geological tim escales, b u t d irect evidence for the ir an tiqu ity has been lacking. H ere, w e describe an un u su a lly dense accum ulation of fossil m arine vertebrates from C erro Ballena, a Late M iocene locality in A tacam a Region of Chile, p reserv ing over 40 skeletons of ro rqual w hales, sperm w hales, seals, aquatic sloths, w alrus-w hales an d p reda to ry b o n y fish. M arine m am m al skeletons are d istrib ­u ted in four discrete ho rizons a t the site, represen ting a recurring accum ulation m echanism . T aphonom ic analysis po in ts to strong spatial focusing w ith a rap id dea th m echan ism a t sea, before being b u ried on a barrier-pro tected sup ra tida l flat. In m o d em settings, HABs are the only know n n a tu ra l cause for such repeated , m ultispecies accum ulations. This p ro p o sed agen t suggests th a t upw elling zones elsew here in the w o rld sh o u ld preserve fossil m arine vertebrate accum ulations in sim ilar m odes a n d densities.

E lectron ic s u p p le m e n ta ry m a te ria l is a v a ila b le a t h t tp : / /d x .d o i .o rg /1 0 .1 0 9 8 /r s p b .2 0 1 3 .3 3 1 6 o r v ia h t tp : / / r s p b .ro y a ls o c ie ty p u b lis h in g .o rg .

Royal Society Pu b lish in g

1. IntroductionD uring th e p as t approxim ately 50 M yr, m arine m am m als evolved in ocean ecosystem s th a t have un d erg o n e g lobal changes in sea level, tem pera tu re , p ro ­duc tiv ity a n d ocean circulation [1 -5 ]. W ith in th is tim e fram e, m u ltip le m arine

© 2 0 14 The A u tho rs. P ub lished by th e Royal Society u n d e r th e te rm s o f th e C reative C om m ons A ttribu tion License h ttp ://c rea tiv eco m m o n s .O rg /lic en se s /b y /3 .0 /, w hich p e rm its u n re s tric ted u se , prov ided th e o rig inal a u th o r a n d source a re cred ited .

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

m am m al lineages evolved from trophic obscurity (i.e. terres­tria l ancestry, w ith little influence on ocean ecosystem s) to ecological dom inance in m arine food w ebs [6-10]. U n d er­stand ing how m arine m am m als, such as cetaceans, p in n ip ed s a n d sirenians, ascended to becom e apex consum ers in m arine food w ebs [11] requires da ta from the fossil record. Palaeobiol- ogists have u sed coun ts of fossil species [2,4,5,12] to outline evo lu tionary changes in richness at the scale of geologic tim e [13], especially d u rin g episodes of m ajor climatic changes [3,4]. These la tter stud ies investigated evo lu tionary causes a n d responses over p ro tracted , diachronic tim e fram es. H ow ­ever, testing ecological in teractions requires d iversity da tase ts from synchronic snapsho ts a t specific scales th a t account for tim e-averaging, sam pling density a n d o ther m etrics of d iver­sity, such as abundance [14]. This la tter goal is a challenge because the m arine m am m al fossil record consists m ostly of sing leton occurrences [13] a n d n o t dense accum ulations.

O b ta in ing ecological sn ap sh o ts o f large m obile p red a to rs such as m arine m am m als is logistically d ifficult because their life-history tra its (e.g. long life, low fecundity , large range) have b ro ad tem poral a n d geographical p aram eters [6 -9]. Palaeoecologists w ork ing w ith terrestria l m am m al an d m arine inverteb rate com m unities have d iscovered tha t sam pling d iversity w ith increased tem poral- an d spatial- averag ing generates dea th assem blage datase ts th a t com pare w ell w ith liv ing com m unities [15-18]. Recently, Pyenson [19,20] d em onstra ted th a t d ea th assem blages o f m o d em ceta­ceans (e.g. strandings) faithfu lly reco rd ecological snapsho ts of liv ing com m unities a t tem poral a n d spatia l scales com ­m ensu ra te w ith the ir m acroecology, w h ich suggests tha t certain fossil assem blages m igh t reta in sim ilarly faithful ecological data . M arine m am m al m ass s trand ings are w ell recorded in historical tim es [19,20], b u t the pu ta tive cases from th e fossil record canno t b e linked to a particu lar causal m echan ism [21,22], especially w ith o u t a b e tte r unders tand ing of the taphonom ic m echanism s th a t b o th preserve a n d p reven t fossil m arine m am m al m ateria l from en tering th e sed im en tary record [23]. H ere, w e describe an u n u su a l accum ulation of fossil m arine vertebrates from th e Late M iocene o f Chile th a t p rov ides un iq ue insights in to the m echanism s tha t p re ­serve dense deposits o f m arine m am m al m aterial, an d the oceanographic processes responsible for their origin.

2. M aterial and m ethods (a ) L oca lityFrom 2010 to 2012, road expansion along the Pan-American Highw ay in Atacam a Region of Chile (figure 1) opened a 20 x 250 m quarry at a site, called Cerro Ballena (27°02'31.51" S, 70°47'42.18" W), w hich revealed over 40 complete and partial m arine m am m al skeletons, along w ith isolated remains of other m arine vertebrates. A road-cut stratigraphie profile of the site exposes approximately 9 m of fine to very fine-grained sandstones belonging to the Bahia Inglesa Formation [24-27], unconformably overlain by a Pleistocene transgressive-regressive m arine terrace sequence [28]. Cerro Ballena is located too far north for strati­graphie correlation w ith other reported localities of the Bahia Inglesa Formation [28], or any recognized mem bers that have been proposed previously ([26] and see the electronic supplem en­tary material). W ithin approximately 8 m of the formation at Cerro Ballena, four different bone-bearing levels (Bone Levels 1 -4 ; BL1 - BL4) produced articulated and associated m arine m am m al fossils (figures 1 and 2). The quarry, w hich is now paved over, represents

Figure 1. Locality a n d g e o g ra p h ic a l in fo rm a tio n fo r Cerro B allena sh o w in g (a) S o u th A m erica w ith (6) th e p a la e o c o a s tlin e o f th e C aldera B asin o u tl in e d over NASA S h u tt le R a d a r T o p o g ra p h y M ission d a ta ; (c) g u a rry m a p sh o w in g sp e c im e n p o s itio n s a n d c o lo u r -c o d e d s tra tig ra p h ie layer, c re a te d in G o o g le E arth ; (d ) v iew o f th e g u a rry a n d th e to p th r e e b o n e -b e a r in g layers ( B L 2 - 4 ) , fa c in g n o r th e a s t , f ro m MPC 6 7 7 in situ. S e e h ttp ://c e rro b a l le n a .s i .e d u a n d th e e le c tro n ic s u p p le m e n ta ry m a te ria l fo r m o re d e ta ils .

only a small portion of the fossiliferous levels, w ith geologic m aps of the unit indicating a local aerial extent of approximately 2 km2 (see the electronic supplem entary material).

(b ) G e o lo g ic a g eThe general region surrounding Caldera contains fossiliferous m arine sedim ents belonging to the Mio-Pliocene age Bahia Inglesa Formation [25], w hich has produced an extensive list of fossil m arine vertebrates [26-28]. We did not find any biostrati- graphically relevant invertebrate and microfossil indicators in strata at Cerro Ballena, although there w ere tw o biostratigra- phically useful vertebrate fossils from Cerro Ballena that also occur in the Mio-Pliocene Pisco Form ation of Peru: aquatic sloths (Thalassocmis) and sharks (Carcharodon). Isolated elements of aquatic sloths from Cerro Ballena are referred to the species Thalassocmis natans [29]. Both T. natans and Carcharodon hastalis from Cerro Ballena are correlated w ith the El Jahuay (ELJ) and

SouthAmerica

Caldera!. - 1 -J - ' \c *-Pajaéoestúarjfra«

view in (d )

□ □ □ □0-50 100-250 250^100 600-700

basement mean height (m)

W70°50 I__

W70°45

rspb.royalsocietypublishing.org Proc. /?. Soc. B

281: 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

,>■

■ OphiomorphaThalassocmis

OphiomorphaThalassinoides, Ophiomorpha

% MNI by BL

Ophiomorpha

Ophiomorpha ■

, • ‘sharks and bony fish

Phocidae

r 'A 'A V

Psilonichnus, Ophiom. Physeteroidea * • . sharks

OdobenocetopsDelphinoidea

Psilonichnus

M I vf f m c vc

grain size

ripple laminationlow-angle planar cross-laminationtrough cross-lamination

ïsh’V r ’

âàl e a ^ ^ B ^

r? = 17

m 77=18

:-v'

j u

Figure 2. S tra tig ra p h ie a n d s e d im e n to lo g ic a l d a ta fro m C erro B a llen a . (a ) U n id e n tif ie d , ir o n -s ta in e d tra c e s re s e m b lin g a lg a l g ro w th s tru c tu re s ; (b) Psilonichnus, a s u p ra tid a l tra c e fo ssil; (c) ir o n -s ta in e d tu f t- lik e fo rm s re s e m b lin g a lg a e , a n d p o s s ib le a lg a l m a ts c o v e rin g w a v e r ip p le s ( in d ic a te d w ith a rro w s) , w ith p e n fo r s c a le a n d {d) s t ra tig ra p h ie c o lu m n , w ith v e r te b r a te d iv e rs ity d a ta e x p re s se d a s p e rc e n ta g e o f th e MNI fro m e a c h b o n e -b e a r in g lay e r (BL1 - 4 ) . {e,f) C rab fe e d in g tra c e s o n th e skull b o n e s o f MPC 6 6 2 , fro m BL-1. c, c o a rse g ra in e d ; EC, E stra to s d e C a ld e ra ; f , f in e g ra in e d ; I, s i l ts to n e ; M, m u d s to n e ; m , m e d iu m g ra in e d ; S, s a n d s to n e ; vc, ve ry c o a rse g ra in e d ; v f, v e ry f in e g ra in e d .

M ontemar Horizon (MTM) horizons in the Sacaco Basin of Peru [30]. In turn, this yields upper and lower bounds on the age of this unit of the Bahia Inglesa Formation at Cerro Ballena as 9.03- 6.45 Ma, following [31] (see the electronic supplementary material). The overlap in stratigraphie range between these two taxa implies that the strata in Cerro Ballena were deposited at a time period in between the deposition of the ELJ and MTM horizons of the Pisco Formation. Thus, w e infer a Late Miocene age (or Late Tortonian to Early Messinian stage) for the Bahia Inglesa Formation part of the section at Cerro Ballena, w hich coincided w ith a rise in sea

level caused by transgressive-regressive cycling and tectonic subsidence along this part of the coastline [32],

(c) D e p o s it io n a l e n v ir o n m e n t a n d s e d im e n t o lo g yIn the road-cut section of the Bahia Inglesa Formation at Cerro Ballena, w e m easured a stratigraphie section, noted sedim entary structures and observed invertebrate trace fossils (figure 2a—c).

Sediment samples collected from BL1 - 4 were freshly excava­ted and covered in optically clear resin (Epo-Tek 301, Epoxy

rspb.royalsocietypublishing.org Proc. R. Soc. B

281; 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

forelimbvertebral column

epiphysealdisks mandible

Figure 3 . O rth o g o n a l d ig ita l th r e e -d im e n s io n a l p o ly g o n m o d e l o f th e m o s t c o m p le te fossil ro rq u a l sp e c im e n a t C erro B a llena , MPC 6 7 7 . T rue n o r th in d ic a te d by a rrow . S e e h t tp ://c e rro b a l le n a .s i .e d u a n d th e e le c tro n ic s u p p le m e n ta ry m a te ria l fo r m o re d e ta ils .

Technology, Billerica, MA, USA). Embedded sam ples were then cut and thin-sectioned for scanning electron microscopy and electron spectroscopy using an FEI Nova NanoSEM 600 under low vacuum w ith the gaseous analytical detector for imaging and an energy dispersive X-ray spectroscopy detector (Thermo- Fisher) for geochemical analysis. Samples were either placed directly on carbon-tape or em bedded in epoxy and thin-sectioned, and left uncoated for SEM and EDS characterization (see the elec­tronic supplem entary material, figure SI). Light microscopy was perform ed using an O lym pus BX51 microscope w ith a Chameleon digital video camera.

(d ) C a p tu r in g , p r o c e s s in g a n d r e n d e r in g th r e e - d im e n s io n a l d ig ita l d a t a s e t s

U nder time-sensitive and salvage circumstances, we docum en­ted in situ skeletal remains using three-dimensional digital tools, before they were collected for study and care at their repositories (see the electronic supplem entary material). Photogramm etry and computer vision datasets for fossil rorquals were captured w ith 20 and 30 cm alum inium scale bars and metal markers (to assess line of sight and control for coverage quality) on a Canon 5D w ith multiple lenses, and geotagged using a Garmin Etrex GPS. We also used Munsell colour charts for colour calibration, accuracy and downstream correction in photography editing software packages. Raw digital datasets were processed into coherent models by aligning datasets, cleaning up noise and removing redundant data using G e o m a g ic v . 2012, P o l y w o r k s

v. 12.0 and Z b r u s h v . 4R3 for the very large dataset of MPC (Museo Paleontológico de Caldera) 677. Direct Dimensions, Inc. (Owings Mills, MD, USA) aligned the laser arm dataset using Geo­magic, and the m odel w as then retopologized using Z b r u s h for MPC 677, creating an orthogonal digital rendering from three- dim ensional polygon data (figure 3). URC Ventures (Redmond, WA, USA) created orthogonal renderings from three-dimensional point cloud datasets (figure 4) by aligning and retopologizing point cloud data (see the electronic supplem entary material, figure S2). The resultant three-dimensional datasets provided sub-centimetre accuracy, and full resolution texture-m apped im agery is available at h ttp ://cerroballena.si.edu .

3. Results and discussionT aphonom ic analysis o f th e site reveals several features tha t are directly com parab le to m o d em m arine m am m al m ass strandings. First, the site p reserves m ultip le species of m arine m am m als, d om ina ted b y ab u n d an t skeletons (M NI = 31;

table 1) of large baleen w hales (clade B alaenopteridae or ro rq­uals) th a t are likely all from the sam e species (see th e electronic su p p lem en ta ry m aterials), a n d encom pass a range of on togen­etic stages, from calves to m atu re ind iv idua ls (figures 3 a n d 4). O ther m arine m am m al species include: (i) a t least tw o different phoc id seals (Acrophoca, an d a new m orphotype); (ii) an extinct species of sperm w hale (Scaldicetus m orphotype); (iii) a w a lru s­like too thed w hale (Odobenocetops) a n d (iv) an aquatic sloth (T. natans) (tables 1 a n d 2; see the electronic supp lem en tary m aterial, figures S3-S7). Second, w e dev ised a sim ple, three- stage categorization to cap tu re the range o f m arine vertebrate taphonom y a t C erro Ballena: (Stage 1) articu lated , either com ­p letely o r m ostly; (Stage 2) d isarticu lated , b u t associated elem ents a n d (Stage 3) isolated, separa ted elem ents. N on­cetacean vertebrates consist of associated, sem i-articulated a n d /o r d isarticu lated skeletal m aterial (Stages 2 a n d 3). By con­trast, ro rqua l skeletons inc luded m an y fully articulated , intact a n d nearly com plete skeletons (Stage 1), a long w ith d isarticu ­la ted skeletons w ith low skeletal scatter (i.e. less th an the d istance of their b o d y length; figure 3; table 2; an d see the elec­tronic supp lem en tary m aterial, tables S l - S l l ) . T hird, rose d iag ram s of th e ro rquals ' skeletons long axis o rien tation (i.e. vertebral colum n) reveal tha t they are orthogonal to curren t flow in each level, analogous to b o d y orien tation pa tte rn s observed for som e m o d em m ass strand ings ([33]; see th e elec­tronic supp lem en tary m aterial, figure S8 an d table S4). Lastly, ro rquals occur m ostly ven tral up , across all BLs (table 2). The dom inance of ven tral u p carcasses, com bined w ith their h igh articulation a n d long axis orientation, is a strong sign tha t they w ash ed in d y ing o r d ead an d w ere then b u ried [34,35].

The dom inance of fossil rorqual skeletons at Cerro Ballena, across all bone-bearing levels, evokes a m odem stranding event, w hereby m any ind iv idual cetaceans are beachcast, either d ead or alive. M ass stranding events for socially gregarious species of toothed w hales are well docum ented [36], b u t m ass strandings for rorquals are rare. The m ost com pelling analogue is a rorqual m ass stranding of 14 hum pback w hales (Megaptera novaeangliae) over the course of five w eeks along approxim ately 50 km of coastline a round C ape Cod, M A, USA in 1987-1988 [35]. This assem blage included males, fem ales an d one calf, w hose necropsies show ed no signs of traum a or predation. Tissue assays of Atlantic mackerel (Scomber scombrus), from stom ach contents, revealed h igh concentrations of saxitoxins, w hich are dinoflagellate neurotoxins. This evidence, along w ith the docum ented aberrant behaviour of one of the dying

rspb.royalsocietypublishing.org Proc. /?. Soc. B

281: 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org o n M arch 14, 2 0 1 4

Figure 4. H igh d y n a m ic ra n g e im a g e s o f o r th o g o n a l th r e e - d im e n s io n a l p o in t c lo u d s c a p tu r in g a d u l t a n d ju v e n ile fossil ro rg u a l s k e le to n s fro m C erro B a llen a , (a) MPC 6 7 8 ; (b) MPC 6 8 4 ; (c) o v e r- la p p in g a d u l t a n d ju v e n ile s p e c im e n s , c lo ck w ise MPC 6 6 6 , 6 6 5 a n d 6 6 7 ; (d ) MPC 6 8 5 a n d (e) MPC 6 7 5 . S m a ll-sc a le b a rs 2 0 c m , la rg e - s c a le b a rs 3 0 c m . T rue n o r th in d ic a te d by a rro w , a n d s tra tig ra p h ie lay e r n o te d b y b o n e -b e a r in g level n u m b e r . S e e h ttp : / /c e r ro b a l le n a .s i .e d u a n d th e e le c tro n ic s u p p le m e n ta ry in fo rm a tio n fo r m o re d e ta i ls a n d so u rc e d a ta .

Table 1. D iversity o f fossil m a r in e v e r te b r a te s a t C erro B a lle n a , w ith m in im u m n u m b e r o f in d iv id u a ls (M NI) b y b o n e -b e a r in g level (BL) a n d w ith ra n g e o f s k e le ta l a r tic u la t io n s ta g e s . This ta b u la t io n d o e s n o t in c lu d e 11 a d d it io n a l , u n id e n t if ie d la rg e c e ta c e a n s k e le to n s (se e th e e le c tro n ic s u p p le m e n ta ry m a te r ia l , f ig u re s S 3 - S 7 a n d S9).

M ysticeti B a la e n o p te r id a e BL 1 - 4 31 S ta g e s 1 - 3

P h o c id ae Acrophoca sp . BL 2 2 S ta g e s 2 a n d 3

E lasm o b ran ch ii Carcharodon hastalis BL 1 , 2 2 S ta g e 3

O d o n to c e ti D e lp h in o id e a BL 1 1 S ta g e 3

O d o n to c e ti P h y se te ro id e a BL 2 1 S ta g e s 2 a n d 3

O d o n to c e ti Odobenocetops sp . BL 1 1 S ta g e 2

P h o c id ae P h o c id a e n . g e n . BL 2 1 S ta g e 3

N o th ro th e r iid a e Thalassocnus natans BL 4 1 S ta g e 3

O s te ic h y th e s Is tio p h o rid a e BL 2 1 S ta g e 3

O s te ic h y th e s X iph iidae BL 2 1 S ta g e 3

rspb.royalsocietypublishing.org Proc. R. Soc. B

281: 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

Table 2. T a p h o n o m ic a t t r ib u te s o f fossil ro rq u a l s k e le to n s a t C erro B a llen a , ra n k e d s tra tig ra p h ic a lly b y BL. N u m b e r o f in d iv id u a l s p e c im e n s (NISP) is sco re d fo r p e rc e n ta g e o r ie n te d v e n tra l u p , s k e le ta l a r tic u la t io n a n d s c a t t e r a n d to ta l le n g th (TL). S e e th e e le c tro n ic s u p p le m e n ta ry m a te r ia l , ta b le s S I — 11.

BL-4 33 3 S ta g e 2 2 .7 7 3 8 .6 2 3

BL-3 67 3 S ta g e 1 2.21 3 7 .63 2

BL-2 67 6 S ta g e s 1 a n d 3 2 .8 0 7 7 .43 7

BL-1 92 12 S ta g e 1 3 .4 5 13 7 .9 7 9

a v e ra g e 75 2 .8 3 7.91

w hales, an d the geographical an d tem poral spans of the event, po in ted to the previously unrecognized trophic transfer of m ajor algal toxins. Since then, o ther cases of harm ful algal b loom s (HABs) involving m arine m am m als have been reported at sim ilar geographical an d tem poral scales [37-41].

The assem blage a t C erro Ballena shares specific features w ith H A B -m ediated m ass strandings. These sim ilarities help delim it the cause of dea th an d th e factors th a t have d riven their concentration a n d p reservation a t th is site. The presence of repeated , m ultispecies assem blages a rgues for a taxonom i- cally b ro ad d ea th m echanism , such as HABs. The proxim ity of m any specim ens, includ ing juvenile an d a d u lt ro rqual skel­etons in d irect contact o r few m etres ap a rt (figure 4c), a long w ith d ifferent m arine vertebrate taxa approxim ately 10 m apart, suggests strong post-m ortem spatia l focusing, p rio r to b u ria l a t each level (figure le). Intraspecific an d interspecific taphonom ic varia tion does n o t elim inate this possibility , as actualistic stud ies of catastrophic cetacean dea th assem blages show a w id e varie ty of decay stages [42]. H A B -m ediated m or­talities a t C erro Ballena w o u ld also p artly explain the absence of vertebrate scavenging a n d the absence of skeletal traum a [43]. The general com pleteness of ro rqual skeletons, in contrast to the d isarticu lation of o ther m arine vertebrates, reflects a size b ias o r tem poral delay in scavenging, p e rm itting m ore d isarti­culation a n d abrasion. Sharks rep resen ted b y iso lated teeth suggest attritional in p u t o r po ten tia l scavenging by-p roducts (see the electronic supp lem en tary m aterial, figure S9). H ow ­ever, billfish rem ains (X iphiidae a n d Istiophoridae) suggest th a t these large p reda to ry consum ers are sim ilarly susceptible to HABs, a find ing th a t h as b een repo rted in th e m o d ern w o rld (see the electronic su p p lem en tary m aterial). Iso lated rem ains of aquatic slo ths (T. natans) m ay reflect incidental, attritional in p u t o r actual H A B -m ediated m orta lity b ased o n extant HAB toxin transfers (i.e. inhalation) for m odern herb ivorous m arine m am m als [40]. Collectively, the tap h o n o m y of Cerro Ballena indicates th a t repeated m arine m am m al m ortalities w ere relatively rap id (hours to w eeks in duration), geograph i­cally w id esp read an d alloch thonous (i.e. a t sea). These latter traits are all consistent w ith H A B -related m ortalities in the m o d em w o rld , w h ich show taphonom ic signals tha t a re tem ­porally delayed a n d physically rem ote from their source [38,40,41,44],

The depositional env ironm ent in w h ich vertebrate car­casses w ere b u ried w as supra tida l, b a sed o n ichnological a n d sedim entological evidence. W e observed a b u n d an t traces of Psilonichnus, w h ich typically occurs o n sup ra tida l flats, a n d Skolithos a n d Ophiomorpha, be longing to the Skolithos ichnofacies, also com m on to tida l flats (figure 2; see the

electronic su pp lem en tary m aterial, figure SIO). In o ther p a rts of the fossil record , Psilonichnus h as been in terp re ted as a crab trace fossil [45]. G iven the un ique food resource p rov ided b y m arine m am m al carcasses, it is n o t su rp ris ing to find scavenging traces on in d iv idua l b a laenop terid bones th a t w e a ttribu te to crabs (figure 2e,f). These sh o rt (approx. 1 cm), sh arp an d closely associated traces o n the skull bones of M PC 662 are sim ilar in trace m orpho logy to those described for pen g u in bones from the M iocene of A rgentina [46]. The lack of varia tion in g ra in size an d scarcity of erosional su r­faces in the section a t C erro Ballena fu rther indicate a m ore or less constan t sed im entation encom passing approxim ately 1 0 -1 6 kyr of deposition , b ased on rates for m o d em tida l flats (see th e electronic supp lem en tary m aterial). The preservation of delicate features resem bling tu fted algal m ats (figure 2c, see [47] an d the electronic supplem entary m aterial, figure S ll) reflects rap id post-m ortem replacem ent b y iron oxides, an d it also indicates the contem poraneous presence of algae and h igh iron concentrations, w hich prom ote algal g row th in this depositional environm ent.

In term s of concentration m echanism , d ead or dy ing m arine vertebrates w ere delivered to south-facing em baym ent, p ro tec ted from norm al w ave action by basem en t rocks an d a ba rrie r b a r to th e w est (figure lb). This in terpre tation is su p ­p o rted b y the absence of n o rth -so u th -o rien ted w ave ripp les (w hich sh o u ld b e p resen t if th is area h a d been directly exposed to Pacific O cean w aves) an d th e presence of low -angle p lan a r cross-bedded sandstone at the to p of the sequence, typical of a beach o r b e rm ( figure 2). S torm su rges flooding the supra tida l flats to a d ep th of ab o u t 1.5 m , as calculated from estim ates of w in d velocity, d u ra tion an d fetch (see the electronic su p ­p lem en tary m aterial), w o u ld have been sufficient to float the largest carcasses from the sou th , allow ing hydrau lic sorting to m odally orien t th em a t each level [48]. The absence of m ajor d isarticu lation a n d lim ited skeletal scatter for any m arine m am m al skeleton fu rther su p p o rts lim ited initial scavenging of floating carcasses, an d rap id transit tim e (hours to days) be tw een dea th a t sea a n d com ing to rest on a pro tected , su p ra tid a l flat. Such a shallow a n d m ostly subaerial env ironm ent also excluded large m arine scavengers. Equally, th e su rro u n d in g desert env ironm ent (w hich already existed a t the tim e) lacked sufficiently large terrestrial p reda to rs [49] th a t cou ld d ism em ber the largest of th e carcasses.

T hus, th e sup ra tida l flat w o rk ed as a taphonom ic trap [19], p reserv ing carcasses th a t a rrived d u rin g sto rm s or sp ring tides in an excellent env ironm ent for decay in situ, m ostly free of scavengers. A s carcasses w ere b u ried u n d e r a relatively con­tinuous rate of fine sed im ents, each m ass s trand ing horizon

rspb.royalsocietypublishing.org Proc. R. Soc. B

281: 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

y ie lded a discrete layer o f skeletal rem ains. Sedim ent sam ples u n d e r ligh t a n d scanning electron m icroscopy lacked distinct algal cell fragm ents (e.g. d ia tom frustules), b u t there w ere w id esp read approxim ately 5 -1 0 |xm spherical apatite g rains encrusted in iron oxides (see the electronic supp lem en tary m aterial, figure SI), w h ich could result from m ineral replace­m en t o f non-siliceous algae (e.g. cyanobacteria) from coastal HABs. H ow ever, w e cannot definitively confirm their b io ­genic source n o r d iscrim inate be tw een their pre-depositional o r post-depositional origins.

A lternative m ass stranding death m echanism s lack m odem analogues or fail to explain the full range of evidence at Cerro Ballena. For exam ple, taxon-specific herd ing , b reeding or strand­ing behaviours do n o t explain the full range of taxa at the site, w hich includes b o th pelagic an d coastal species th a t do no t inha­b it supratidal environm ents (table 1). Tsunam is w o u ld have generated death assem blages lacking large b o d y size selectivity (tables 1 a n d 2) an d w o u ld result in h igh-energy sedim entary structures tha t are no t present. Pandem ic causes, such as m orbillivirus, are n o t taxonom ically b road , n o r w o u ld it be parsim onious as a recurring m echanism over 10 -1 6 kyr. Thus, all o ther alternative death m echanism s besides HABs fail to explain the iterative p reservation o f four bone-bearing levels.

The excavation quarry a t Cerro Ballena y ie lded a density of associated, fossil m arine m am m al skeletons un riva led else­w here in the w orld . The density of ind iv idua l cetacean specim ens a t C erro Ballena, for exam ple, is greater th an o ther a ttritional deposits in th e cetacean fossil record, including th e Shark too th H ill bonebed from th e M idd le M iocene of C alifornia [23] an d th e Eocene lagoonal deposits of W adi A l-H itan in E gypt [50]. The density of ro rqual skeletons in BL1 (see th e electronic supp lem en tary m aterial, table S12) alone is 10 tim es greater th an associated, ind iv idua l densi­ties repo rted from the M io-Pliocene Pisco Basin of sou thern P eru [51], w h ich are p reserved in distal, m arine shelf env iron­m ents, a n d no tab ly lack m ultip le m arine m am m al species in close association (less than 10 m). C erro Ballena is also su r­p rising in its abundance of com plete ro rqual skeletons because baleen w h ale s trand ings are com paratively rare in th e m o d ern w orld . W e p ropose th a t th is ecological asym m etry arises from th e shifted baseline o f ba leen w hale abundances. In rem ote areas today, such as th e S ou thern O cean, there are exam ples of super-aggregations th a t are u n u su a l by to d ay 's s tandards, b u t m atch historical a n d anecdotal accounts of baleen w h ale abundances p rio r to industria l w haling [52].

Evidence for H A B -m ediated dea th assem blages of m arine vertebrates in the fossil record is lim ited to only a few cases because of th e difficulty in a ttribu ting HABs as a causal agen t [53,54]. M o d em analogues of m arine m am m al deaths caused by HABs ou tline a likely p a th w ay tha t occurred

repeated ly a t th is site d u rin g the Late M iocene. W e p ropose |th a t toxins, generated b y HABs, po isoned m ultip le species of m arine vertebrates, th rough ingestion of con tam inated prey a n d /o r inhalation , causing relatively rap id dea th a t sea. C arcasses th en floated tow ards th e coastline, w here they en tered the estuary an d w ere tran spo rted b y locally generated , n o rth w ard p ropaga ting sto rm w aves into a restricted sup ra ti­d a l flat, w here they w ere b u ried to the exclusion of m ajor scavenging a n d d isarticulation . This sequence w as recor­d e d four tim es d u rin g th e deposition of sed im en t (approx. 1 0 -1 6 kyr) at C erro Ballena. The conditions th a t lead to this repeated phenom enon are tied to upw elling system s along w esterly m arg ins of continental coastlines [55]. A long the coast o f w estern South A m erica, ferrug inous runoff from the A ndes leads to increased iron in th e ocean, w h ich boosts p ro ­ductiv ity w here iron is a lim iting n u tr ien t for phy to p lan k to n g row th [55], w hile also p rom o ting HABs (e.g. cyanobacteria a n d dinoflagellates [44]). The an tiqu ity of these processes p robab ly p re-dates m o d ern tectonic configurations, a lthough it h as n o t b een d ocum en ted in th e fossil record un til now . W e p ropose th a t upw elling system s, fuelled b y iron-rich runoff, in o ther regions of the w o rld w ill have sim ilar, repeated accum ulations of m arine consum ers [56].

A c k n o w led g em en ts . W e th an k tw o an o n y m o u s review ers for com m ents th a t im proved th e m anuscrip t. For d iscussion a n d access to unpub lish ed data , w e th an k P. D. G ingerich, R. E. Fordyce, D. E. G. Briggs, J. R. Geraci, G. A. Early, L. M azzuca, J. Castle, J. H . R odger Jr, J. G. M ead, C. W. Potter, D. J. B ohaska, D. R. L indberg , J. A. G oldbogen, M . T. C lem entz, S. L. W ing a n d A . K. B ehrensm eyer. For assistance in th e field, w e th an k C. P. F igueroa-B ravo, R. E. Y ury-Y añez, S. Fuentes Tam blay, S. H illebrandt, D. Tom ini, R. Terreros, J. A revalo, D. del M onte an d B. Tornini. W e are in d eb ted to G. Kovak-Lewis, B. Sem erjian an d D. B oardm an a t URC V entures for analy tical support. A lso, w e th an k D. Cole (Sm ithsonian CIS C oordinator), J. C ope (N F orm ation D esign), E. L. M on tano for assistance w ith N A SA elevation datasets, T. D zam bazova (A utodesk), A . V elevski an d R. W iedem eier for w ebsite assistance, P. Fu (3D Systems), C. G om ez (M useo N acional d e H isto ria N atu ra l in Santiago), L. López an d the Consejo d e M o n um en tos N acionales (CM N), an d C. W aibel, J. Blundell, an d L. Lew is of th e Sm ithson ian 's D igitization P rogram Office for their su p p o rt. This p ap e r is C aldera Paleontology Project con tribu tion no . 4.

Funding S ta tem en t. Excavation w o rk w as co n d u c ted u n d e r C M N p e rm it no. 5979 to M.E.S. a n d w as financed b y Sacyr C hile S.A. F un d in g from CONICYT, Becas Chile, D ep artam en to d e P ostg rado y P ostítu lo of th e V icerrectoría d e A su n to s A cadém icos of U n iv ersid ad d e C hile su p p o rte d C.S.G. This w o rk w as also fu n d e d b y a N atio n a l M u seu m of N a tu ra l H isto ry (N M N H ) Sm all G ran t A w ard , d isc re tionary fu n d ­in g from N M N H Office of th e D irector, th e S m ithson ian In stitu tio n 's R em ing ton K ellogg F und , tw o N atio n a l G eographic Society C o m m it­tee o n R esearch E xploration g ran ts (8903-11, 9019-11) to N.D.P. a n d b y U-REDES (D om eyko II U R -C 12 /1 , U n iv e rs id ad d e Chile) to A. O. V argas.

References

1. Norris RD, K irtland T urner S, Hull PM, Ridgw ell A.2013 M arine ecosystem resp o n ses to Cenozoic g lo b a l ch a n g e . Science 3 4 1 , 492 - 498 . (do i:10 . 1 1 2 6 /sc ie n ce .l 240543)

2. Lipps JH, M itchell Jr ED. 1976 Trophic m o d e l for th e a d a p tiv e rad ia tio ns an d ex tin c tion s o f pe lag ic m a rin e m a m m a ls . Paleobiology 2 , 4. 1 4 7 - 1 5 5 .

3 . Fordyce RE. 2003 Cetacea ev o lu tio n an d E o c e n e -O lig o c e n e o cean s rev isited . In From greenhouse to icehouse. The marine Eocene- Oligocene transition (ed s DR P ro thero , LC Ivany,EA N esb itt), p p . 1 5 4 - 1 7 0 . New York, NY: C olum bia University Press.M arx FG, U hen MD. 2 010 C lim ate, c ritte rs , an d c e taceans: Cenozoic d rivers o f th e ev o lu tio n o f

m o d e rn w h a le s . Science 3 2 7 , 9 9 3 - 9 9 6 . (do i:10 . 1 1 2 6 /sc ie n ce .l 185581)

5. Su to I, K aw am ura K, H ag im o to S, Teraishi A, Tanaka Y. 2 012 C hanges in upw ellin g m ech an ism s d rove th e evo lu tion o f m a rin e o rg an ism s. Palaeogeogr. Palaeoclimatol. Palaeoecol.3 3 9 - 3 4 1 , 3 9 - 5 1 . (do i:10 .1 0 1 6 /j.p a la e o . 2012 .04 .014 )

rspb.royalsocietypublishing.org Proc. R. Soc. B

281: 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

6. Schw arz LK, G oebel ME, Costa DP, Kilpatrick AM. 2013 T op-dow n a n d b o tto m -u p in fluences on d em o g rap h ic ra te s o f A ntarctic fu r sea ls Arctocephalus gazella. J. Animal Ecol. 82, 9 0 3 - 9 1 1 . (doi :10 .1111 / 1 365 -26 5 6 .1 2 0 5 9)

7 . S p rin g e r AM, Estes JA, van V liet GB, W illiam s TM, D oak DF, D anner EM, Forney KA, Pfister B. 2003 S eq u en tia l m e g a fa u n a l co llap se in th e N orth Pacific O cean: a n o n go ing legacy o f industria l w h a ling ? Proc. Natl Acad. Sei. USA 100, 12 2 2 3 - 1 2 228. (do i:10 .1 0 7 3 /p n a s . l 6 3 5156100 )

8. Ainley DG et al. 2 010 Im pac ts o f c e tacean s on th e s tru c tu re o f S o u the rn O cean food w eb s . Mar. Mamm. Sei. 26, 482 - 498 . (d o i:1 0 .1 1 1 1 /j.1 74 8 - 7 6 92 .2009 .00337 .x )

9 . Croll DA, M arinovic B, B enson S, Chavez FP, Black N, Ternullo R, Tershy BR. 2005 From w ind to w h a le s : tro p h ic links in a coasta l upw ellin g sy stem . Mar. Ecol. Prog. Ser. 289, 1 1 7 - 1 3 0 . (d o i:1 0 .3 3 5 4 / m ep s2 8 91 1 7 )

10. V e lez -Juarbe J, D om ning DP, P yenson ND. 2012 Iterative ev o lu tion o f sym patric seacow (D u g o n g id ae , S irenia) a sse m b la g e s d u rin g th e p a s t ~ 2 6 m illion years. PLoS ONE 7, e3 12 9 4 . (do i:10 . 1 3 7 1 /jo u rn a l.p o n e .0 0 3 1 2 9 4 )

11. T erbough J, Estes JA (ed s). 2 010 Trophic cascades: predators, prey, and the changing dynamics o f nature. W a sh in g to n , DC: Island Press.

12 . M arx FG. 2 009 M arine m a m m a ls th ro u g h tim e : w h e n less is m o re in th e s tu d y o f p a laeod iversity . Proc. R. Soc. B 276, 8 8 7 - 8 9 2 . (do i:1 0 .1 0 9 8 /rsp b . 2 008 .1473 )

13. U hen MD, Pyenson ND. 2 0 07 Diversity e s tim a te s , b iases, an d h isto rio graph ic e ffec ts: resolving c e tacean diversity in th e T ertiary. Palaeontol. Electrónica. 10, 11 A.

14. P yenson ND, Irm is RB, Fipps JH. 2 010 C o m m en t on 'C lim ate , c ritte rs , a n d ce taceans: Cenozoic d rivers o f th e evo lu tion o f m o d e rn w h a le s '. Science 330, 178. (do i:10 .1 1 2 6 /sc ie n c e .l 189866)

15. Terry RC. 2 009 The d e a d do n o t lie: u sing skele ta l rem a in s fo r rapid a s se s sm e n t o f historical sm all- m a m m a l c o m m u n ity base lin e s . Proc. R. Soc. B 277,

1193 - 1 2 0 1 . (do i:10 .1 0 9 8 /rsp b .2 0 0 9 .1 984)16. M iller JFH. 2011 G hosts o f Y ellow stone: m u lti-

decad a l h istories o f w ild life p o p u la tio n s c a p tu re d by b o n e s on a m o d e rn lan d scap e . PLoS ONE 6 , e18057 . (d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 1 80 5 7 )

17. Kidwell SM. 2001 P rese rva tion o f sp ec ies a b u n d a n c e in m a rin e d e a th a sse m b la g e s . Science 294,

1091 - 1 0 9 4 . (do i:10 .1 1 2 6 /sc ie n c e .l 064539)18. Kidwell SM. 2 002 T im e-av eraged m o lluscan d e a th

a sse m b la g e s : p a lim p ses ts o f richness, sn a p sh o ts o f a b u n d a n c e . Geology 30, 8 0 3 - 8 0 6 . (d o i:1 0 .1 1 3 0 / 0 0 91 -7 6 1 3 (2 00 2 )0 3 0 < 0803:TAMDAP > 2.0.CO;2)

19. P yenson ND. 2 010 C arcasses o n th e coast: m easu rin g th e eco logical fide lity o f th e ce tacean s tran d in g record in e a s te rn N orth Pacific O cean. Paleobiology 3 6 , 4 5 3 - 4 8 0 . (do i:10 .1 6 6 6 /0 90 1 8 .1 )

20 . Pyenson ND. 2011 The high fidelity o f th e cetacean strand ing record: insights in to m easuring diversity by in tegra ting tap h o n o m y an d m acroecology. Proc. R. Soc. B 278, 3 6 0 8 - 3 6 1 6 . (doi:10.10 9 8 /rspb .2011 .0441 )

2 1 . K em per CM, P led g e N, Fing JK. 1997 Subfossil ev id en ce o f s tran d in g s o f th e sp e rm w h a le Physeter macrocephalus in G ulf St V incent, S ou th A ustralia. Rec. South Australia Mus. 29, 4 1 - 5 3 .

2 2 . B a ldanza A, Bizzarri R, F am ian i F, M onaco P, P elleg rino R, Sassi P. 2013 E nigm atic, b iogenically in d u ced s truc tu res in P le is to cen e m a rin e d ep o sits : a first record o f fossil a m b e rg ris . Geology 41,

1 0 7 5 - 1 0 7 8 . (do i:10 .1130 /G 34731 .1)2 3 . P yenson ND, Irm is RB, B arnes FG, M itchell J r ED,

M cFeod SA, Fipps JH. 2 009 Origin o f a w idesp read m a rin e b o n e b e d d e p o s ite d d u rin g th e m id d le M iocene C lim atic O p tim u m . Geology 37, 5 1 9 - 5 2 2 . (do i:10 .1130 /G 25509A .1)

2 4 . M arq u ard t C, Favenu A, O rtlieb F, Godoy E, C om te D. 20 04 Coastal n eo tec ton ics in S o u thern Central A ndes: up lift a n d d efo rm atio n o f m a rine te rraces in N orthern Chile (27°S ). Tectonophysics 394, 193 - 21 9 . (doi:10 . 1016 /j .te c to .2004 .07 .05 9)

2 5 . A churra FE, Facassie JP, Fe Roux JP, M arq u ard t C, B elm ar M, R uiz-del-S olar J, Ishm an SE. 2009 M a n g an ese n o d u les in th e M iocene Bahia Inglesa F o rm ation , n o r th -cen tra l Chile: p e tro g rap hy , geoch em is try , g en esis a n d p a laeo cean o g rap h ic sign ificance . Sed. Geo!. 217, 1 2 8 - 1 3 9 . (do i:10 .1 0 1 6 /j.se d g e o .2 0 0 9 .0 3 .0 1 6)

2 6 . W alsh SA, S uárez ME. 2 006 N ew p en g u in rem ains from th e P liocene o f n o rth e rn Chile. Hist. Biol. 18,

115 - 1 2 6 . (do i:10 .1 0 8 0 /0 8 9 1 2 9 60 6 0 0 64 0 7 9 6 )2 7 . G utstein CS, Cozzuol MA, V argas AO, Suárez ME, Schultz

CF, Rubilar-Rogers D. 2009 P attern s o f skull variation o f Brachydelphis (Cetacea, Odontoceti) from th e N eogene o f th e S o utheastern Pacific. J. Mamm. 90, 5 0 4 - 5 1 9 . (doi:10.1644/07-M A M M -A -081.1 )

2 8 . V alenzuela-T oro AM, G utste in CS, Varas-M alea RM, Suárez ME, Pyenson ND. 2013 P inn iped tu rn o v e r in th e S o u th Pacific O cean: n ew ev id en ce from th e P lio -P le istocene o f th e A tacam a D esert, Chile.J. Vert. Paleontol. 33, 2 1 6 - 2 3 3 . (d o i:1 0 .1 0 8 0 / 0 2 72 4 6 3 4 .2 01 2 .7 1 0 2 82 )

2 9 . M uizon C, d e M cD onald HG. 1995 An a q u a tic slo th from th e P liocene o f Peru . Nature 375, 2 2 4 - 2 2 7 . (d o i:1 0 .1 0 3 8 /3 7 52 2 4 a0 )

3 0 . M uizon C, M cDonald HG, Salas R, U rbina M. 2 004 The y o u n g es t species o f th e aq ua tic slo th Thalassocnus an d a re a sse sm e n t o f th e re la tionsh ip s o f th e n o th ro th e re s lo th s (M am m alia : X enarth ra). J. Vert. Paleontol. 24, 3 8 7 - 3 9 7 . (d o i:10 .1671 /2429a)

3 1 . E hret DJ, M acfadden BJ, Jo n e s DS, Devries TJ, Foster DA, Salas-G ism ondi R. 2 012 Origin o f th e w h ite sh a rk Carcharodon (F am n ifo rm es: F am n id ae) based on recalib ra tion o f th e U pper N eo g ene Pisco F orm ation o f P eru . Palaeontology 5 5 , 1 1 3 9 - 1 1 5 3 .

3 2 . Fe Roux JP et al. 2005 N eo g ene-Q ua te rn ary coasta l an d o ffsho re s e d im e n ta tio n in n o rth - cen tra l Chile: record o f sea -lev e l c h a n g e s an d im p lica tions fo r A ndean te c to n ism . J. South Am. Earth Sei. 19, 8 3 - 9 8 . (d o i:1 0 .1 0 1 6 /j.jsam es.20 0 3 .11 .003)

3 3 . Geraci JR, S t. Aubin DJ. 1977 M ass s tran d in g o f th e lo n g -fin n ed p ilo t w h a le , Globicephala melaena, on S ab le Island, Nova Scotia . J. Eish. Res. Board Can. 34, 2193 - 21 96 . (do i:1 0 .1 1 3 9 /f7 7 -2 8 8)

34 . S chäfer W. 1972 Ecology and palaeoecology o fmarine environments. C hicago, IF: University o f Chicago Press.

35 . Geraci JR, A nderson DM, T im peri RJ, St. Aubin DJ, Early GA, P resco tt JH, Mayo CA. 1989 H um pback w h a le s (Megaptera novaeangliae) fa ta lly po isoned by d in o fla g e lla te to x in . Can. J. Eish. Aguat. Sei. 46,

1 8 9 5 - 1 8 9 8 . (do i:1 0 .1 1 3 9 /f8 9 -2 3 8 )36 . S e rg ea n t DE. 1982 M ass s tran d in g s o f to o th e d

w h a le s (O don toceti) a s a p o pu la tio n p h en o m e n o n . Sei. Rep. Whales Res. Inst. Tokyo 34, 1 - 4 7 .

37 . Van Dolah FM, D o u ce tte GJ, G ulland F, B ossart G, Row les T. 2003 Im pac ts o f a lgal to x in s on m a rin e m a m m a ls . In Toxicology o f marine m ammals (ed s J Vos, GD B ossart, M Fournier, T O 'Shea), p p . 2 4 7 - 2 6 9 . F ondon , UK: Taylor & Francis.

38 . d e la Riva GT, Jo h nson CK, G ulland FM, F anglois GW, H eyning JE, Row les TK, M aze t JA. 2009 A ssociation o f an u n usu a l m a rin e m am m al m o rta lity e v e n t w ith Pseudo-nitzschia s p p . b loom s a lo n g th e so u th e rn California co astline . J. Wildi. Dis. 45, 1 0 9 - 1 2 1 . (d o i:1 0 .7 5 8 9 /0 0 90 -3 5 5 8 -4 5 .1 .1 0 9 )

39 . Scholin CA et al. 2 000 M orta lity o f sea lions a long th e cen tra l California co ast linked to a tox ic d ia to m b lo o m . Nature 403, 8 0 - 8 4 . (do i:1 0 .1 0 3 8 /4 7 48 1 )

40 . Flew elling FJ et al. 2005 Red t id e s a n d m arine m a m m a l m o rta litie s . Nature 435, 7 5 5 - 7 5 6 . (d o i:1 0 .1 0 3 8 /n a tu re4 3 5 7 5 5 a)

41 . Fire SE, W ang Z, B erm an M, Fanglois GW, M orton SF, Sekula-W ood E, Benitez-Nelson CR. 2010 Trophic transfer o f th e harm ful algal toxin dom oic acid as a cause o f d ea th in a m inke w ha le (Balaenoptera acutorostrata) stranding in sou thern California. Aguat. Mamm. 36,

3 4 2 - 3 5 0 . (doi:10.1578/AM .36.4.2010.342)42 . F iebig PM, Flessa KW, Taylor SA. 2 007 T aphonom ic

varia tion d e sp ite ca tas trop h ic m o rta lity : analy sis o f a m ass s tran d in g o f fa lse killer w h a le s (Pseudorca crassidens), G ulf o f C alifornia, M exico. Palaios 22,

3 8 4 - 3 9 1 . (do i:10 .2 1 1 0 /p a lo .2 0 0 5 .p 0 5 -0 5 2 r)43 . Fallow s C, G allagher AJ, H am m ersch lag N. 2013

W hite shark s (Carcharodon carcharias) scaveng ing on w h a le s a n d its p o te n tia l ro le in fu r th e r sh ap in g th e eco logy o f a n ap ex p red a to r.PLoS ONE 8, e6 0 7 9 7 . (d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e . 0060797)

44 . Silver MW, Bargu S, Coale SF, B en itez-N elson CR, Garcia AC, R oberts KJ, S eku la-W ood E, B ruland KW, Coale KH. 2 0 10 Toxic d ia to m s a n d do m o ic acid in na tu ra l an d iron en rich ed w a te rs o f th e ocean ic Pacific. Proc. Natl Acad. Sei. USA 109, 20 7 6 2 -20 7 6 7 . (d o i:1 0 .1 0 7 3 /p n as .1 0 06 9 6 8 10 7 )

45 . Frey RW, Curran HA, P e m b erto n SG. 1994 T racem aking activ ities o f crabs a n d th e ir e n v iro n m en ta l s ign ificance: th e ich n o gen u s Psilonichnus. J. Paleontol. 58, 3 3 3 - 3 5 0 .

46 . C ione AE, H o sp ita leche CA, Perez FM, Faza JH, Cesar I. 2 010 Trace fossils on p e n g u in b o n e s from th e M iocene o f C h u b u t, so u th e rn A rgen tina .Alcheringa 34, 433 - 454 . (d o i:1 0 .1 0 8 0 / 03115 5 1 10 0 3 7 93 4 7 0 )

47 . M ayall M, W righ t VP. 1981 Algal tu f t s truc tu res from th e u p p e r Triassic o f so u th w e s t E ngland. Palaeontology 24, 6 5 5 - 6 6 0 .

rspb.royalsocietypublishing.org Proc. /?. Soc. B

281: 20133316

D ow n load ed from rsp b .roya lsoc ietyp u b lish in g .org on M arch 14, 2 0 1 4

48 . R ogers RR, Kidwell SM. 2 007 T he orig in an d in te rp re ta tio n o f b o n e b e d s . In Bonebeds: genesis, analysis, and paleobiological significance (ed s RR R ogers, DA E berth,AR Fiorillo), pp . 1 - 6 4 . C hicago, IL: University o f Chicago Press.

49 . Prevosti FJ, Forasiepi AM, Zimicz N. 2013 The ev o lu tion o f t h e Cenozoic te rres tria l m am m a lia n p re d a to r gu ild in S o u th A m erica: c o m p etitio n or rep la c e m e n t? J. Mamm. Evol. 2 0 , 3 - 2 1 . (do i:10 .10 0 7 /s i 0 9 14 -0 1 1 -9 1 75 -9 )

50 . G ingerich PD. 1992 M arine m a m m a ls (C etacea an d Sirenia) from th e Eocene o f Gebel M okattama n d Fayum , E gypt: s tra tig rap h y , a g e , an d

p a le o e n v iro n m e n ts . Univ. Mich. Papers Paleont.3 0 , 1 - 8 4 .

5 1 . Brand FR, E speran te R, C hadw ick AV, P om a 0 ,A lom ia M. 2 0 04 Fossil w h a le p rese rva tion im plies h igh d ia to m accum ula tio n ra te in th e M io c e n e - P liocene Pisco F orm ation o f Peru . Geology 3 2 ,1 6 5 - 1 6 8 . (do i:10 .1130 /G 20079 .1)

5 2 . N ow acek DP, F ried laen d er AS, FHalpin PN, FHazen EF, Jo h n s to n DW, Read AJ, E spinasse B, Zhou M, Zhu V.2011 S u p e r-ag g re g a tio n s o f krill a n d h u m p back w h a le s in W ilhelm ina Bay, A ntarctic P en in su la . PLoS ONE 6 , e1 9 1 7 3 . (d o i:1 0 .1 3 7 1 /jo u rn a l.p o n e .0 0 1 91 7 3 )

5 3 . Em slie SD, M organ GS. 1 994 A ca ta s trop h ic d e a th 56. a s se m b la g e a n d p a le o d im a tic im p lica tions o f

Pliocene seab ird s o f F lorida. Science 2 6 4 , 6 8 4 - 685 . |(do i:10 .1 1 2 6 /sc ien ce .2 6 4 .5 1 59 .684)

54 . C astle JW , R odgers Jr JH. 2 009 H ypo thesis fo r th e ro le o f to x in -p ro d u cin g a lg a e in P hanerozo ic m ass ex tin c tion s b ased on ev id en ce from th e geo log ic record a n d m o d e rn en v iro n m e n ts . Environ. Geosci. 1 6 , 1 - 2 3 . (d o i:1 0 .1 3 0 6 /e g .0 8 1 10808003)

55 . D ezileau F, Ulloa 0 , H ebbeln D, Fam y F, Reyss J-F, F o n tu g n e M. 2 0 04 Iron contro l o f p a s t productiv ity in th e coasta l upw ellin g system o ff th e A tacam a D esert, Chile. Paleoceanography 1 9 , PA3012.(doi :10 .1 0 2 9 /2004P A 001006)B ron g e rsm a-S an d ers M. 1957 M ass m o rta lity in th e sea . Geo!. 5oc. Am. Mem. 1 6 7 , 9 4 1 - 1 0 1 0 .

rspb.royalsocietypublishing.org Proc. R. Soc. B

281: 20133316