Evolution of the Naticid Gastropod Predator-Prey System: an Evaluation of the Hypothesis of Escalation Patricia H

Western Washington University Western CEDAR

Geology Faculty Publications Geology

8-1993 Evolution of the Naticid Gastropod Predator-Prey System: An Evaluation of the Hypothesis of Escalation Patricia H. Kelley

Thor A. Hansen Western Washington University, [email protected]

Follow this and additional works at: https://cedar.wwu.edu/geology_facpubs Part of the Geology Commons

Recommended Citation Kelley, Patricia H. and Hansen, Thor A., "Evolution of the Naticid Gastropod Predator-Prey System: An Evaluation of the Hypothesis of Escalation" (1993). Geology Faculty Publications. 3. https://cedar.wwu.edu/geology_facpubs/3

This Article is brought to you for free and open access by the Geology at Western CEDAR. It has been accepted for inclusion in Geology Faculty Publications by an authorized administrator of Western CEDAR. For more information, please contact [email protected]. 358 RESEARCH REPORTS

Evolutionof the Naticid Gastropod Predator-PreySystem: An Evaluation of the Hypothesisof Escalation

PATRICIA H. KELLEY

Departmentof Geologyand GeologicalEngineering, University of NorthDakota, Grand Forks,ND 58202

THOR A. HANSEN

Departmentof Geology,Western Washington University, Bellingham, WA 98225

PALAIOS, 1993, V. 8, p. 358-375 factorsincreasingly have been viewedas importantin evolution (Van Valen, 1973; Vermeij,1987). Predation and are biologichazards to whichspecies Previous workhas suggested that escalation may have competition major either extinction,or adaptation. characterizedthe historyof the naticid gastropodpred- respond, by migration, Interactionsamong competitors or betweenpredator and ator-preysystem, based on apparent increasesin drilling preymay providesignificant input into the evolutionary frequenciesand the occurrenceof antipredatoryapta- tions amongprey. We evaluate this hypothesisbased on process. (1987) has hypothesizedthat a dominanttheme a comprehensivesurvey (over 40,000specimens) pre- Vermeij of of evolution the Phanerozoichas been escalation, dation on molluscs the Cretaceous through during from Upper to enemies. by one spe- lower withinthe U.S. and At- involvingadaptation Adaptation Oligoceneformations Gulf to have a deleteriouseffect on other lantic Coastal Plain. cies tends species withinthat effective environment; therefore, species must Patternsin drilling bothbivalve and gastropodprey of to in the same place" (Van Valen's are complex.Drilling frequencies were relativelylow in adapt simply "stay "Red 1973).As a result,Vermeij (1987, the Cretaceous but increased sharply above the Creta- Queen hypothesis," has "the historyof lifeis characterizedby ceous-Tertiaryboundary, remaining high until the late p. 4) suggested, two simultaneoustrends: increasing risks to individuals Eocene. Followinga significantdecline near the Eocene- from enemiesand increasingincidence and ex- Oligocene boundary,drilling frequencies increased to a potential of to cope with these risks. Modern moderatelevel in the Oligocene.Contrary to our predic- pression aptations be no betteradapted to their biological tion based on the hypothesis escalation, no temporal organismsmay of than ancient ones were to theirs,but the trend increasingstereotypy drillholesite occurred. surroundings of of have themselvesbecome more rig- However,significant increases in prey effectiveness(in- biologicalsurroundings orous withina habitat."The idea that biologichaz- dicated bythe incidence of incomplete drillholes and mul- given ards have become increasinglysevere, while aptationsto tiplybored shells) occurredbetween the Cretaceousand such hazards have become betterexpressed, has been re- Oligocene. This pattern characterizesentire faunas as ferredto as the "hypothesisof escalation." well as individualprey taxa that wereconsistently heavi- by Vermeij This was criticizedby Kitchell (1990, p. 154), ly drilled (turritellidgastropods and corbulidbivalves). hypothesis whoclaimed "it emphasizesthe expectationof progressive or linear trendsin aptationsand hazards and omitsany explicitconsideration of the consequencesof tradeoffsor INTRODUCTION feedback,"and thatit assumesa constantadaptive "value" fora giventrait. These criticismsappear to be unfounded; The relativeimportance of bioticand abioticfactors in Vermeij (1987) extensivelydiscussed adaptive tradeoffs evolutionhas been the subject of much debate in evolu- and indicatedthat thereneed not be a tightcorrespon- tionarybiology (see Ross and Allmon,1990). Changes in dence betweenthe expressionof a traitand the functional the physical environmenthave long been recognizedas demandsplaced on it. Nevertheless,the hypothesisof es- agentsof natural selection. In recentyears, however, biotic calation remainscontroversial. Vermeij (1987) discussed

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 359

possible argumentsagainst this hypothesis,including: 1) fectivenesswould be especially favoredunder such cir- nonbiologicalfactors may be more importantagents of cumstances.In addition,since naticidsselect preybased naturalselection than are biologicalfactors; 2) adaptation on morphology,rather than at random,it mightbe ex- may be preventedby such factorsas geneticor develop- pectedthat prey would respond by evolvingantipredatory mentalconstraints and tradeoffsamong aptations; and, 3) morphologicaptations. selectionat the level of the individualmight be opposed Some previouswork supports the hypothesisof escala- to or less importantthan selectionat otherlevels. tion forthe naticid predator-preysystem. Allmon et al. The purpose of the present study is to examine the (1990) foundthat drillingfrequencies were significantly historyof a particularpredator-prey system, that of drill- lowerfor Cretaceous species of turritelline gastropods than ing naticid gastropodsand theirmolluscan prey, and to forPaleocene, Miocene,or Plio-Pleistocenespecies. Data apply these resultsto evaluatingthe hypothesisof esca- cited by Vermeij (1987) suggestthat drillingfrequencies lation. We have undertakena comprehensivesurvey of formolluscs were low duringthe Cretaceousbut mayhave naticid predationfrom 14 stratigraphiclevels withinthe reachedmodern levels by the Eocene (see also Dudley and Cretaceousand Paleogene of the U.S. Gulf and Atlantic Vermeij,1978; Vermeijet al., 1980). Thus, the biological Coastal Plain. More than 40,000specimens have been ex- hazard of naticidpredation appears to have becomemore amined in this survey,most from bulk samples. severe,at least fromthe Cretaceousto sometimein the Paleogene. However,assessment of trendsin drillingfre- THE PREDATOR-PREY SYSTEM quencies based on previousstudies is difficult.The data presentedby Vermeij include only two data points for The Naticidae are one of severalfamilies of shell-drill- Cretaceousthrough Paleocene gastropodassemblages, and ing,predatory gastropods. Naticids, which are shallowin- onlyone pointfor bivalves. Allmon et al. (1990) also noted faunalsnails, have been a majorsource of molluscanmor- the inadequacyof available Cretaceousdata sets. talitysince the Cretaceous (Sohl, 1969). As summarized In addition to possible increases in the frequencyof by Carriker(1981), preyare envelopedwithin the preda- drilling,naticids may have becomemore efficient predators tor's large mesopodiumand orientedin a preferredposi- throughtime. Previous studies indicate that drillholepo- tion fordrilling. Drilling, using the radula and secretions sitionmay have been morevariable in the Cretaceousand of the accessoryboring organ, may take severalhours to Paleogene than in the Neogene (comparestudies of Ade- days. Ingestionis by means of the proboscis,which is in- goke and Tevesz, 1974; and Kitchell,1986, withthose of sertedthrough the drillhole. Berg and Nishenko,1975; Hoffmanand Martinell,1984; Studies of extantnaticids (Kitchell,1986; Boggs et al., and Kelley, 1988, 1991,among others). Adegoke and Te- 1984) indicatethat drillingbehavior is highlystereotyped vesz (1974) also reporteda lack of preysize selectivityin in that prey are not evaluated based on experiencesof an Eocene assemblagefrom Nigeria, while Neogene sam- individualnaticids; predator behavior involves long-term ples generallyshow size selectivity(Kitchell et al., 1981; responsesdeveloped over evolutionarytime. Because of Kelley, 1988, 1991; Hoffmanand Martinell,1984; Kabat this, it is possible to study naticid behaviorusing time- and Kohn, 1986). In addition,mean size ofspecies of three averagedfossil assemblages (Kitchell, 1986). Stereotyped naticid generaapparently increased during the Cenozoic behaviorof extant naticids includes selectivity of drillhole (Kitchell,1986). Because largerpredators are able to at- site,prey size, and preyspecies. Such selectivitymaximizes tack largerprey, such size change would have increased energygain per foragingtime: prey selection in laboratory the hazard of naticidpredation. trialswas predictableby cost-benefitmodelling (Kitchell Vermeij (1987) argued that, in response to increased et al., 1981). Cost to the predatoris determinedby drilling hazards of predation,antipredatory aptations have be- time,which is controlledby preyshell thickness.Benefit come morepronounced. For instance,global surveyssug- is a functionof biomass,which is highlycorrelated with gestan increasein thefrequency of mobile or well-armored the internalvolume of the shell.Within a preyspecies, the gastropodsand ofbivalves with tight valve closure through cost: benefit(thickness: internal volume) ratio generally geologictime (Vermeij,1987). In addition to the global decreaseswith increasing prey size. Predatorschoose prey surveys,evidence of increasedantipredatory aptations ex- itemswith the lowestcost: benefitratio in a size rangethat ists for individuallineages (Kelley, 1989). Thickness in- can be handled. creaseswere documented within five common Chesapeake A plausible hypothesisis that escalationmay have oc- Group bivalve genera,apparently in responseto naticid curredwithin the naticidpredator-prey system over geo- predationduring the middleMiocene. Such thicknessin- logic time. Naticids mightbe expectedto have improved creaseseffectively deterred predation, as drillingfrequen- predatorycapabilities for several possible reasons: 1) in cies withina taxondecreased as shellthickness increased. responseto prey throughcoevolution, which is the evo- Althoughstudies of naticid predationexist forassem- lution of two species in response to one another (see blages representinga varietyof ages and paleoenviron- DeAngeliset al., in press); 2) due to competitionbetween ments,the historyof naticid predationcannot be recon- naticidpredators (Vermeij, 1987); or, 3) by predationon structedfrom the publishedliterature. As Vermeij(1987, the predatoritself (Vermeij, 1983). In the case of naticids, p. 423) stressed,"the hypothesisof escalationcan be eval- whichare cannibalistic,potential naticid prey are also po- uated properlyonly in temporalcomparisons of similar tentialpredators and competitors.Increased predator ef- nonbiologicalenvironments." In addition,previous studies

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 360 KELLEY & HANSEN

TABLE 1 -Stratigraphic levels employed in the study. Formationsand members are listed in descending stratigraphicorder.

Formation Age Locality

Byram EarlyOligocene Mississippi MintSpring EarlyOligocene Mississippi Red Bluff EarlyOligocene Mississippi MoodysBranch Late Eocene Louisiana,Mississippi Gosport Late middleEocene Alabama Cook Mountain* Late middleEocene Texas, Louisiana UpperLisbon* Late middleEocene Alabama PineyPoint* Late middleEocene Virginia Bashi** EarlyEocene Alabama Bells Landing** Late Paleocene Alabama MatthewsLanding** MiddlePaleocene Alabama Brightseat EarlyPaleocene Maryland Kincaid EarlyPaleocene Texas Corsicana Late Cretaceous Texas Providence Late Cretaceous Georgia Ripley Late Cretaceous Georgia

* The CookMountain, Upper Lisbon and Piney Point Formations are roughly synchronous and are referred to in thetext as the"Cook Mountainlevel." **Bashi, Bells Landing, and Matthews Landing are members of the Hatchetigbee, Tuscahoma, and Naheola Formations respectively. vary greatlyin the methodsused and types of data re- the Ripley and Providence Formations,housed at the ported.Many wereconducted prior to the developmentof U.S.G.S. in Reston,Virginia, were made available by Nor- an appropriatemodel of naticidpredation (Kitchell et al., man Sohl. The sample ofthe MatthewsLanding Member 1981), producingenigmatic and sometimescontradictory of the Naheola Formationwas providedby L.W. Ward of results. theVirginia Museum of Natural History. All othersamples To circumventthese problems,we have conducted a were collectedby the authors.Those fromthe Brightseat comprehensive,systematic study of the naticidgastropod Formationare housedat the U.S. NationalMuseum; sam- predator-preysystem in the U.S. Gulfand AtlanticCoast- ples fromthe remainingunits are at WesternWashington al Plain. If the hypothesisof escalation is valid, naticid University. predator-preyinteractions may have differedearly in the The assemblageswere composed almost entirely of mol- historyof the groupcompared to later.We investigatethis luscs, mostlybivalves and gastropods.In all samples,the idea by a historicalapproach, concentrating on the largely originalshell materialwas preserved,and there was no unstudied(and putativelycritical) early history of naticid evidence of chemicaldissolution. The species diversities predation (Cretaceous and Paleogene). Following Ver- ofthe samplesare generallyhigh and the shellsdisplay no meij's suggestion,our data are from"similar nonbiological physicalevidence of taphonomicalteration, such as shell environments"represented by a temporalsuccession of 14 wear or size sorting. stratigraphiclevels. This comprehensivesurvey allows di- All ofthe collectionsfrom the BrightseatFormation and rect comparisonof predation on different-agedassem- stratigraphicallyhigher units were preserved well enough blages and testingof hypotheseson the historyof preda- to wet sieve and treatwith the followingprocess. After a tion. small bulk sample was washed and dried,all shells were picked, identified(using the taxonomyof Palmer and Brann,1965, 1966, with modifications and additionsfrom MATERIALS AND METHODS Bretsky,1976; Dockery, 1977,1982; MacNeil and Dockery, Materials 1984; and Stephenson,1941) and counted,thereby pre- servingthe relativeabundances of species. Of thissample, Our samples werederived from 16 formationscompris- only those specimensthat were relativelycomplete were ing 14 stratigraphiclevels (single formations or, in the case observedand countedfor naticid predation. Shell recovery of the Cook Mountain level, 3 roughlycontemporaneous was good forsamples in the Brightseatand above,and we formations,the Piney Point, upper Lisbon, and Cook believe that the naturalrelative abundances, albeit time- Mountain) withinthe Atlanticand Gulf Coastal Plain, averaged,of most species were well representedin our rangingfrom the Upper CretaceousRipley Formationto washed collectionswith the exceptionof some large,thin- theOligocene Byram Formation (Table 1). Collectionsfrom shelledspecies, e.g., some species of Tellinidae. These sam-

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 361

pies thus providevaluable informationon overalldrilling as one with at least 10 specimens(20 valves, in the case rates and prey preferencesof naticidsthrough time and of bivalves). in differentassemblages. Data on drillholemorphology and positionwere tabu- Shells fromthe Kincaid and CorsicanaFormations were lated foreach drilledspecimen. Drillhole shape indicates not sufficientlywell preservedto wash throughsieves, so whetherthe predatorwas a naticidor muricidgastropod. they were carefullycleaned with needles, leaving them Naticid drillholescharacteristically are bevelled,and par- exposed on matrixpedestals. Only the exterioror interior abolic in cross section;muricid drillholes are unbevelled of each bivalve valve and only one face of a gastropod and possess straightsides. This studyfocusses on preda- could be observed.Therefore only bivalve exteriorswere tion by naticids.Drillhole site on bivalveswas recorded countedfor drilling frequencies, because gastropodscould withrespect to a nine-sectorgrid previously used byKelley not be exposedsufficiently to ensurecomplete observation (1988, fig.1). Drillholesite was recordedfor gastropods as of the shell, and examinationof bivalve interiorswould well,but we have not yet analyzed these data forstereo- not reveal incompletedrills or drillholemorphology. typy.In addition,we recordedwhether each hole was com- The Ripley and Providencespecimens were fromcol- plete (entirelypenetrating the shell) or incomplete,indi- lectionsthat had been made overmany years by Norman cating an unsuccessfulpredation attempt.We did not Sohl and others.Therefore these collections represent un- recognizecomplete but nonfunctional drillholes. Such holes usually completespecies assemblages,but we cannot be have inner:outer borehole diameter ratios less thanabout confidentthat the relativeabundance of each species has 0.5 (Kitchell et al., 1986), indicatingthat the inner di- been maintained.For these samples,overall drilling fre- ameterwas not enlargedenough to allow passage of the quencies are susceptibleto bias, but drillingfrequencies proboscis.Because of the magnitudeof the study,it was withinspecies should be accurate. impracticalto measureborehole diameters. Most of the data points representtwo or more strati- graphicallyseparate samples fromeach formation,taken at the same locality,which were then combined to provide RESULTS an average forthat formation.The Cook Mountain and Spatial and EnvironmentalPatterns in Moodys Branch levels were much more widelysampled, DrillingFrequencies includinga broad spectrumof lithologies,biofacies, and regions.These samplestherefore provide an averagebased Little is knownabout the environmentalvariability of on a much widercollection. Analysis of variationswithin naticid predationin modernoceans. Drillingfrequencies these levels providesan importantindicator of the range fromall sources (includingnon-naticid drillers) seem to of drillingfrequencies found at one time under different be higherin the tropicsthan in temperateregions (Ver- environmentalcircumstances (Hansen and Kelley,in prep.). meij, 1987; thoughVermeij et al., 1989,present evidence All ofthe sampledformations represent normal marine, fora possible reversedlatitudinal gradient of drillingon innerto middle shelf,clastic sedimentaryenvironments. bivalves). There is some evidencethat drillingfrequency Grain size varies fromsilty clay to glauconiticsand. The increaseswith water depth (Sander and Lalli, 1982). Drill- shallowestwater samples representinner shelf, sandy en- ing predationfrom all sources varies considerablyin a vironmentswell below normal wave base. The deepestwa- region,from 5 to 38% of the fauna (Vermeij,1987). ter samples representmiddle shelfmuddy environments Studieson thegeographic and environmentalvariability (the CretaceousCorsicana Formationrepresents a some- of naticiddrilling in fossilassemblages are virtuallynon- what deepermid-outer shelf environment). In generalthe existent.Typically, data on fossildrilling frequencies rep- climatevaried from warm temperate to subtropical(Wolfe resenta sample fromone locality,not reflectingpossible and Poore, 1982). environmentalvariations. An exceptionis the study by Allmonet al. (1990), whichexamined roughly contempo- Data Tabulation raneous populationsof Turritellafrom Virginia and Al- abama. In this case the Virginiaspecimens of Turritella Drillingfrequencies were tabulated for each bivalveand were foundto have higherdrilling frequencies, contrary gastropodtaxon withineach stratigraphiclevel. Drilling to the modernlatitudinal trend. It is importantto assess frequencywas definedas D/N forgastropods and 2D/N howvariable drilling frequency can be at one timein order for bivalves,where D = numberof successfullydrilled to evaluate the significanceof observedtemporal trends. specimensand N = totalnumber of specimens; "specimen" Comparedto the otherstratigraphic levels in our study, signifiesa singlevalve in the case of bivalves.In the same the Cook Mountain and Moodys Branch levels represent way, overall drillingfrequencies were calculated for the the largestsample size and the broadestspectrum of en- total bivalveand gastropodfaunas of each level. Two ad- vironments.The Cook Mountain level includes samples ditionalmeasures of predationintensity were determined fromthe sandymud of the PineyPoint Formationin Vir- foreach level: 1) thepercent of species successfully drilled; ginia,muddy sand of the upper Lisbon Formationof Al- and, 2) the percentof abundantspecies withdrilling fre- abama, and the mostlymud Cook MountainFormation of quencies greaterthan 0.10. The lattermethod was used Louisiana and Texas. Frequencyof drillingfor bivalves by Vermeij (1987) as a measure of predationintensity; and gastropods(combined) fromindividual assemblages followinghis convention,we definean abundant species of the Cook Mountain level varies from7-39%, with a

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 362 KELLEY & HANSEN

mean of 21%. This level spans a greaterlatitudinal and The Moodys BranchFormation has no appreciablelat- climaticrange than does anyother level in ourstudy (warm itudinalrange, but does include a varietyof sedimentary temperateto subtropical).But althoughthere is apprecia- facies,from shelly glauconitic sand to mostlymud. Drilling ble environmentalvariation within this level,drilling fre- frequenciesfor the Moodys Branch level range from4- quency shows no directcorrelation with these variables. 22%, with a mean of 10%. The sample containing22% Drillingfrequency does not correlatewith sediment size, drilledindividuals is somewhatanomalous; all othersam- distancefrom shore, percentage of bivalves,or percentage ples have drillingfrequencies less than or equal to 13%. of naticids(Hansen and Kelley,in prep.). As in the case of the Cook Mountain level, drillingfre- quencyin the MoodysBranch does not correlatewith sed- imentsize or distancefrom shore. The wide rangeof drillingfrequencies found within one stratigraphiclevel suggests that predation rates from small or geographicallyrestricted samples must be regardedas minimumvalues.

Q4 a Temporal Patternsin DrillingFrequencies 5 ...... | .. Vermeij (1987) suggestedthat a temporalincrease in I|I ^drilling frequencieswould indicate intensificationof the ...... hazard of predationand thus escalation.His global com- j...|j pilationssuggest that drilling frequencies were low during I ||___ the Cretaceous,reaching modern levels by the Eocene. The C BR BL CM 3 MS realityand timingof such an increase has been contro- PR K ML BS G RB BY versial (Allmonet al., 1990). STRATIGRAPHICIEVEL Our data indicatethat naticid predation on bivalvesand gastropodscombined was fairlylow forUpper Cretaceous |I BIVALVES _ GASTROPODS | TOTAL samples and rose abruptlyin the middle Danian (Paleo- cene) samplesto essentiallymodern levels (over30 %, Fig. FIGURE 1 -Drilling frequenciesthrough time for bi% alves, gastropods, 1). Extremelylow levels of predationoccur in both the and totalfauna. Abbreviationsfor stratigraphic levE als: RP, Ripley;PR, uppermostCretaceous Corsicana and lowermostPaleocene Providence; C, Corsicana; K, Kincaid; BR, Brightseat; ML,Matthews Kincaid but these data are somewhat Landing; BL, Bells Landing; BS, Bashi; CM, Cook Mountain;G, Gos- Formations, points port; MB, Moodys Branch; RB, Red Bluff;MS, MintSpring; BY, suspect because of low sample sizes and relatively poor Byram.Axis not scaled to indicatethe stratigraphic spacing of levels. preservation. High predation levels were maintained

TABLE 2-Numbers ofspecies (sp.) and specimens (spec.), drillingfrequency (dr. freq. = 2 x numberof successfully drilledspecimens: total numberof specimens), and percentof species successfullydrilled (% sp. dr.) forall bivalve species and infaunalbivalve species withineach formation.Also shown is the percentof abundant bivalvespecies (i.e., those withmore than 20 specimens) that had drillingfrequency >0.10 (% ab. sp. dr.).

All species Infaunalspecies # of # of Dr. % Sp. % Ab. # of # of Dr. % Sp. Level sp. spec. freq. dr. sp. dr. sp. spec. freq. dr.

Ripley 18 2629 0.132 72.2 61.6 17 2211 0.156 70.6 Providence 19 639 0.194 52.6 83.3 19 639 0.194 52.6 Corsicana 37 291 0.055 16.2 0 26 209 0.077 23.1 Kincaid 43 1054 0.034 18.6 7.7 38 855 0.042 21.0 Brightseat 45 1743 0.327 37.8 81.8 40 1635 0.347 40.0 MatthewsLanding 8 235 0.111 62.5 33.3 7 210 0.124 71.4 Bells Landing 15 62 0 0- 13 42 0 0 Bashi 13 1124 0.415 61.5 100.0 13 1124 0.415 61.5 Cook Mountain 90 5035 0.287 46.7 74.2 75 4627 0.289 52.0 Gosport 30 468 0.073 23.3 20.0 25 454 0.070 24.0 MoodysBranch 54 13,372 0.086 63.0 36.7 50 12,881 0.089 62.0 Red Bluff 31 2197 0.235 19.4 50.0 23 2031 0.254 26.0 MintSpring 38 1596 0.204 57.9 76.9 31 1505 0.210 64.5 Byram 20 662 0.139 45.0 100.0 18 654 0.141 50.0

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 363

throughthe Paleocene and early to middle Eocene and widelyduring the Cretaceousthrough Oligocene. Preda- then dropped abruptlyto Cretaceous levels in the late tion on the bivalve fauna as a whole closelyreflects that middleand late Eocene. Predationrates rose again in the measured forinfaunal bivalves only,as infaunalspecies earlyOligocene, but did not attain the highlevels of the dominateall the samples. Paleocene and earlyEocene. Drillingfrequency was moderatelyhigh in the Creta- Naticid gastropodpredation on bivalvesis summarized ceous Ripleyand ProvidenceFormations in termsof per- in Table 2 and Figures1 and 2. Predationon bivalves,as cent of bivalveindividuals drilled, and veryhigh in terms indicatedby all threemethods of measurement(% of in- ofpercent of species that experienced drilling. In addition, dividuals,species, and abundant species drilled),varied a large percentof the abundant species in the fauna ex- hibited drillingfrequencies greater than 10% (Table 2; Fig. 2). All measuresof predationthen declined signifi- 0 5 AL -W I- to low values in the Cretaceous 6 1 cantly very uppermost 7 _ Corsicana Formationand lowermostPaleocene Kincaid A 0.9- 6 11 Formation;drilling frequencies are significantlydifferent = z 0.8- 31 -3_ forthe Providenceand Corsicana(x2 13.89,P < 0.001). c 0.7- Bivalvesfrom the lower Paleocene BrightseatFormation 13 6 15 6 experiencedrelatively high predation; 33 % of individuals 0.6- A, 10 and 38% ofspecies were drilled, and 82 % ofthe abundant z 0.5- had than 10%. This 6 species drillingfrequencies greater increaseis highlysignificant (x2 = 145.80with 1 degreeof , 0.4 3 freedomfor a comparisonof Kincaid and Brightseatdrill- 0.3- 3 ing frequencies).Although the percentof bivalve species 5 z 0.2- drilledwas largein the middlePaleocene MatthewsLand- 5 .. . 11 weredrilled. The < 0.1 . inglevel, only % ofbivalve individuals 23 uppermostPaleocene Bells Landing sample is dominated o 0 I I w I I I by gastropods,and none ofthe fifteenbivalve species pres- t> RP PRBR C BR ML L BS CM G MB RB MS BY STRATIGRAPHICLEVEL ent was drilled. Drilling frequencywas substantialin both the lower A I* BIVALVES GASTROPODS | Eocene Bashi and middle Eocene Cook Mountain levels (42 % and 29% forindividuals, respectively). The percent FIGURE 2-Percent of abundant species with drillingfrequencies ofindividuals drilled then dropped significantly (x2 = 42.12, greater than 0.10. Abundant species are those with more than 20 with 1 of to 7-9% in the and valves (bivalves) or 10 specimens (gastropods). Numberof species degree freedom), Gosport indicatedabove (bivalves) or below (gastropods) each data point.Ab- Moodys Branch.(In contrast,63 % ofthe Moodys Branch breviationsfor stratigraphiclevels as in Figure 1. Axis not scaled to bivalvespecies were drilled, though at verylow levels. This indicatethe stratigraphicspacing of levels. is probablyan effectof the large sample size forthe Moodys

TABLE 3-Numbers of species (sp.) and specimens (spec.), drillingfrequency (dr. freq. = numberof successfully drilledspecimens: totalnumber of specimens), and percentof species successfullydrilled (% sp. dr.)for all gastropod species and infaunalgastropod species withineach formation.Also shown is the percentof abundant gastropod species (i.e., those withmore than 10 specimens) that had drillingfrequency >0.10 (% ab. sp. dr.).

All species Infaunalspecies # of # of Dr. % Sp. % Ab. # of # of Dr. % Sp. Level sp. spec. freq. dr. sp. dr. sp. spec. freq. dr.

Ripley 29 2426 0.037 51.7 8.7 12 1484 0.032 58.3 Providence 11 516 0.060 54.5 20.0 4 312 0.083 100.0 Brightseat 18 414 0.377 66.7 50.0 11 275 0.484 72.7 MatthewsLanding 20 297 0.465 50.0 100.0 5 181 0.486 80.0 Bells Landing 26 742 0.352 38.5 60.0 14 692 0.366 50.0 Bashi 32 271 0.203 43.8 66.7 12 204 0.191 50.0 Cook Mountain 59 2761 0.162 44.1 45.5 28 2441 0.175 50.0 Gosport 23 200 0.075 21.7 33.3 10 171 0.088 50.0 MoodysBranch 46 938 0.109 41.3 37.5 13 476 0.099 61.5 Red Bluff 39 329 0.119 23.1 20.0 13 172 0.134 15.4 MintSpring 51 737 0.145 52.9 66.7 23 552 0.111 56.5 Byram 25 178 0.174 48.0 66.7 9 70 0.200 55.6

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 364 KELLEY & HANSEN

TABLE 4-Distribution of boreholes with respect to a nine-sectorgrid, and results of chi-squared tests. DMV, dorsal-middle-ventralcomparison; AMP, anterior-middle-posteriorcomparison; 9S, nine sector comparison. Ital- icized chi-squared values are not significantat the 0.05 level. All othervalues are significantat least at the 0.05 level.

Sector Chi-squared Level Taxon 1 2 3 4 5 6 7 8 9 DMV AMP 9S

Ripley Striarcasp. 0 0 0 0 3 0 2 16 0 26.6 31.1 Caestocorbulasp. 1 6 1 2 17 3 0 1 0 22.1 27.2 Corbulasp. 3 6 1 3 19 2 4 7 1 7.5 28.4 49.1 Lucina parva 1 0 0 4 8 0 1 1 0 14.8 8.4 Providence Striarcacuneata 0 2 1 1 11 0 0 3 0 9 25 Corbulasp. 0 2 0 1 7 3 3 5 2 6.3 7.9 Brightseat Pitar sp. 0 5 1 2 7 0 0 0 0 8.4 14.8 Corbulasubengonata 0 4 2 3 5 0 2 3 1 0.4 6.7 Crassatellaalaeformis 0 7 5 4 23 3 2 2 0 23.1 27.3 78.4 Vokesulaaldrichi 1 1 1 1 6 2 3 6 3 5.2 4.8 Venericardiaplanicosta 1 3 2 5 6 1 0 1 0 9.6 3.9 Venericardiasp. 1 10 3 4 9 5 1 4 2 4.8 12.2 Phacoidessp. 2 10 0 8 63 2 1 8 0 83.3 119.4 309 Bashi Venericardiahoratiana 1 0 1 12 26 12 32 31 18 71.4 7.6 88.6 Corbulasubengonata 3 0 1 3 12 3 6 1 3 9.3 1.9 Vokesulaaldrichi 0 0 2 1 17 15 3 8 1 31.9 14.6 66.6 Cook Mountain Lucina pomilia 4 12 2 3 33 4 1 3 0 31.9 54.3 125 Spisula parilis 2 8 2 3 8 4 1 1 2 6.3 6.6 Callistaperovata 4 12 2 3 3 3 0 1 0 15.5 7.4 Caryocorbuladeusseni 1 2 2 4 4 3 1 6 2 2.2 2.5 Corbulaextenuata 1 2 0 3 9 9 2 6 1 15.3 5.6 Pachecoa decisa 6 35 3 4 20 1 3 10 4 13.4 69.5 104 MoodysBranch Lucina subcurta 2 15 4 4 28 8 2 4 4 19.5 35.9 74.1 Lucina curta 3 8 0 6 12 3 3 1 0 12.2 13.5 Caestocorbulawailesiana 1 3 2 9 17 2 3 8 7 14 9.9 36.3 Spisula jacksonensis 2 7 5 3 8 5 0 4 1 5.9 8.5 Red Bluff Scapharca invidiosa 0 0 1 2 20 8 1 2 2 41.2 15.2 Astartetriangulata 0 0 2 5 14 4 4 4 6 17.1 3.2 Corbularufaripa 7 9 6 19 50 18 21 29 21 38.2 19.6 73.7 MintSpring Corbulalaqueata 3 1 4 4 18 10 11 16 14 21.6 5.4 34.4 Byram Corbulalaqueata 1 1 1 3 8 2 0 3 3 7.2 4.7

Branch.) Predation in the three lower Oligoceneforma- differencesare statisticallysignificant; for instance, a com- tions occurredat levels comparableto those in the Cre- parisonof Providenceand Brightseatdrilling frequencies taceous. These values are significantlyhigher than those producesa x2 of 143.44 with 1 degreeof freedom.Signif- forthe upper Eocene (x2 = 206.44, Moodys Branch and icantlylower drilling frequencies occurred in all theEocene Red Bluffcomparison). and Oligocene levels; percentof individualsdrilled was Predationintensity on gastropods(Table 3, Figs. 1, 2) less than or equal to 20%. The percentof species drilled was lowin the Cretaceous,high in thePaleocene and some- fluctuatedduring this interval between about 20 and 50%; what reducedin the Eocene and Oligocene.Although epi- these values are lowerthan those forthe lowerPaleocene faunalgastropods were abundant in thesamples, predation samples. on infaunalgastropods was similarto that forthe gastropod fauna as a whole.Stereotypy of Naticid Behavior Althoughabout halfthe Cretaceous(Ripley and Prov- idence)gastropod species were drilled, very few individuals The dynamicsof the naticid-preysystem may have dif- weredrilled (4% and 6% forthe twoformations), and the feredduring the historyof the group. If the hazard of percent of abundant species with drillingfrequencies naticidpredation increased as predictedby the hypothesis greaterthan 10% was also relativelylow. All measuresof of escalation,we suggestthat naticid behaviormay have predation on gastropods were high in the Paleocene been less stereotypedearly in theirhistory. We investi- Brightseat,Matthews Landing, and Bells Landing. The gatedone aspectof naticid behavioral stereotypy, drillhole

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICID GASTROPOD ESCALATION 365

TABLE 5-Numbers and percentages of all bivalve species, successfullydrilled bivalve species, infaunalbivalve species (inf.sp.), and infaunalsuccessfully drilled bivalve species (inf.dr. sp.) bearing incompleteand multiple holes.

All species Drilledsp. Inf.sp. Inf.dr. sp. # w/ # w/ % w/ % w/ % w/ % w/ % w/ % w/ % w/ % w/ Level inc. mul. inc. mul. inc. mul. inc. mul. inc. mul.

Ripley 3 0 16.7 0 23.1 0 17.6 0 25.0 0 Providence 1 0 5.3 0 10.0 0 5.3 0 10.0 0 Corsicana 2 0 5.4 0 33.3 0 3.8 0 16.7 0 Kincaid 1 1 2.3 2.3 12.5 12.5 2.6 2.6 12.5 12.5 Brightseat 3 3 6.7 6.7 17.6 17.6 7.5 7.5 18.8 18.8 MatthewsLanding 3 1 37.5 12.5 60.0 20.0 42.9 14.3 60.0 20.0 Bells Landing 1 0 6.7 0 - 7.7 0-- Bashi 2 3 15.3 23.1 25.0 37.5 15.4 23.1 25.0 37.5 Cook Mountain 19 7 21.1 7.8 45.2 16.7 20.0 8.0 38.5 15.4 Gosport 2 1 6.7 3.3 28.6 14.3 8.0 4.0 33.3 16.7 MoodysBranch 15 12 27.8 22.2 44.1 35.3 30.0 24.0 48.4 38.7 Red Bluff 10 5 32.3 16.1 166.7 83.3 39.1 17.4 150.0 66.7 MintSpring 4 2 10.5 5.3 18.2 9.1 12.9 6.5 20.0 10.0 Byram 2 1 10.0 5.0 22.2 11.1 11.1 5.6 22.2 11.1

siteselectivity. Such selectivitypromotes drilling the thin- selectivityof drillholesite. One of three Bashi species, nest area of the shell (Kitchell,1986) and correlateswith Corbulasubengonata, did not showdrillhole selectivity in greaterprobabilities of successful predation (Kelley, 1988). the anterior/middle/posteriorcomparison; another spe- If naticidswere less capable earlyin theirhistory, drillhole cies, Venericardiahoratiana, had a broad distributionof site maybe morevariable for taxa fromolder assemblages. drillholesdespite rejectionof the null hypothesis.In the Boreholedistributions for all taxa weretreated with two Cook MountainFormation, the null hypothesiscould not chi-squaredtests. The firsttested whether dorsal (sectors be rejectedfor Corbula extenuata (anterior/middle/pos- 1-3), middle (sectors4-6), and ventral(sectors 7-9) areas teriorcomparison) and Caryocorbuladeusseni (bothcom- ofthe shell were equally drilled. The secondtested whether parisons). Equal drillingof dorsal, middle, and ventral anterior(sectors 1, 4, and 7), middle (sectors2, 5, and 8), regionscould not be rejectedfor Spisula jacksonensis of and posterior(sectors 3, 6, and 9) areas of the shell were the Moodys Branch Formation. equally drilled.Eleven species had sufficientnumbers of Only fourOligocene species had sufficientnumbers of drillholesthat chi-squaredtests could be used to test the drillholesfor statistical analysis of their distribution. Cor- nullhypothesis of equal drillingof each ofthe nine sectors. bula laqueata (Mint Springand ByramFormations) and Table 4 providesdata on the distributionof drillholes Astarte triangulata (Red Bluff)showed no anterior-to- forbivalve taxa from9 ofthe 14 levelsstudied. (The other posteriorpreference for drilling. Boreholes were broadly fivelevels containedno bivalvespecies drilledfrequently distributedon Corbula rufaripa,but with a statistically enough to allow statisticalanalysis of boreholedistribu- significantpreference for sector 5; Scapharca invidiosa tion.)No generaltemporal trend in stereotypyis apparent. was also preferentiallydrilled in the valve center. All of the six Cretaceousspecies in the data set showed of boreholesite < 0.05 for selectivity (P all chi-squared Relative Effectivenessof Predatorand Prey tests). Five of the six species displayeda preponderance of boreholesin mid-valve(sector 5), while boreholesin Two typesof data wereused to determinewhether the Striarca fromthe Ripley Formationtended to occurin a relativeeffectiveness of predator and preychanged through mid-ventralposition (sector 8). Somewhatless stereotypy time: the incidenceof incompletedrillholes and that of is evidentfor the Paleocene BrightseatFormation. Of the multiplydrilled shells. Incompletedrillholes result from seventaxa withsufficiently abundant drillholes, only three unsuccessfulpredation attempts; following Vermeij (1987), yieldedsignificant chi-squared values forall thetests. The we used the ratioof incompleteto attempteddrillholes to null hypothesiscould not be rejectedfor Corbula suben- indicate"prey effectiveness." In addition,the presenceof gonata and Venericardiasp. (dorsal/middle/ventralcom- multipleboreholes in a shell (normallyone functionalhole parison),nor for Venericardiaplanicosta (anterior/mid- and one or more unsuccessfulholes, but sometimesmul- dle/posteriorcomparison) or for Vokesula aldrichi (both tiple completeholes) may indicatesuccessful prey escape comparisons). tactics (Kitchellet al., 1986). Most,but notall, frequentlydrilled Eocene taxa showed We consideredfirst the percentageof species in a for-

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 366 KELLEY & HANSEN

TABLE 6-Numbers of complete (C.D.) and incomplete(inc.) drillholesand preyeffectiveness (P.E. = numberof incompleteholes:total attemptedholes) forbivalve species that exhibitedincomplete drilling.

Level Species C.D. Inc. P.E.

Ripley Striarcasp. 21 2 0.087 Caestocorbulasp. 31 2 0.061 Corbulasp. 46 6 0.115 Providence Striarcacuneata 14 1 0.067 Corsicana Ostrea 0 2 1.0 Corbulasp. 1 1 0.5 Kincaid Saccella sp. 0 1 1.0 Brightseat Corbulasubengonata 21 15 0.417 Vokesulaaldrichi 24 11 0.314 Cucullaea gigantea 5 1 0.167 Matthews Landing Cucullaea macrodonta 7 1 0.125 Venericardiawilcoxensis 2 1 0.333 Corbulasubcompressa 2 2 0.5 Bells Landing Vokesulaaldrichi 0 1 1.0 Bashi Corbulasubengonata 32 7 0.179 Vokesulaaldrichi 47 4 0.078 Cook Mountain Nuculana multilineata 3 1 0.25 Saccella catasarca 3 1 0.25 Pachecoa decisa 266 5 0.018 Pachecoa declivis 10 2 0.167 Limopsisaviculoides 12 1 0.076 Chlamysclarkeana 1 1 0.5 Eburneopectenscintillatus 0 1 1.0 Cubiotstreasellaeformis 54 6 0.10 Plicatula filamentosa 1 1 0.5 Lucina pomilia 151 2 0.013 Venericardiacf. V. alticostata 7 4 0.364 Bathytormussp. 5 2 0.286 Spisula parilis 32 2 0.059 Pitar sp. 2 1 0.333 Anapterisregalis 1 1 0.5 Corbulacf. C. alabamiensis 5 1 0.167 Corbulaextenuata 40 5 0.111 Corbulasubengonata 10 1 0.091 Caryocorbuladeusseni 28 10 0.263 Gosport Lyrodiscustellinoides 0 1 1.0 cf.Caestocorbula murchisonii 2 1 0.333 Moodys Branch Nucula spheniopsis 15 3 0.167 Limopsisradiata 5 5 0.500 Lucina curta 56 5 0.082 Lucina subcurta 100 4 0.038 Venericardiadiversidentata 18 3 0.143 Pleuromerisinflatior jacksonensis 11 1 0.083 Lirodiscusjacksonensis 1 1 0.5 Bathytormusflexurus 6 3 0.333 Crassinellapygmaea 3 1 0.25 Spisula jacksonensis 37 4 0.098 Alveinusminutus 169 56 0.249 Kelliella boettgeri 31 15 0.326 Callista annexa 3 2 0.4 Corbuladensata 27 8 0.229 Caestocorbulawailesiana 66 16 0.195 Red Bluff Barbatiaparadigona 0 1 1.0 Scapharca (S.) invidiosa 36 3 0.077 Limopsis sp. 0 1 1.0

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 367

TABLE 6-Continued.

Level Species C.D. Inc. P.E.

Glycymerissp. 0 1 1.0 Eburneopectensubminutus 0 5 1.0 Venericardiacarsonensis 0 2 1.0 Timothynusturgida? 3 1 0.25 Astartetriangulata 41 3 0.068 Corbula(V.) rufaripa 180 34 0.159 Corbula(Caryocorbula) engonata 3 9 0.75 MintSpring Ervilia lamelloexteria 1 1 0.5 Corbula(Caryocorbula) engonata 14 5 0.263 Corbula(Varicorbula) laqueata 81 17 0.173 Corbularufaripa 4 2 0.333 Byram Corbula(Caryocorbula) engonata 1 1 0.5 Corbula(Varicorbula) laqueata 22 7 0.241

mationthat exhibitedincomplete drillholes. Although the frequencyof bivalve species with incompletedrillholes TABLE 7-Data fromTables 6 and 10, summarizedby fluctuatedthrough time, there is some evidenceof a tem- family.(Only familieswith infaunal species listed.) Also poral increasein incompletedrilling (Table 5). With the shown, foreach family,is the numberof species with exceptionof the Ripley and MatthewsLanding levels, few- incompleteholes. (Species occurringat more than one er than 7 % ofthe speciesin the Cretaceousand Paleocene stratigraphiclevel are counted morethan once.) Families containedincomplete drillholes. In contrast,of the Eocene listed in order of preyeffectiveness. and Oligoceneunits, only in the GosportFormation did fewerthan 10% of the species contain incompletedrillholes. About 30% of the in the No. species present Moodys of Branch and Red BluffFormations contained incomplete drillholes.The is whetherall bivalve spe- pattern similar, prey Family C.D. Inc. P.E. cies species or only infaunalspecies are considered.The patternis notsimply an artifactof sample size; thecorrelation Bivalves betweennumber of specimensin a sample and percentof 1 1.0 1 = Glycymeridae 0 species with incompletedrillholes is nonsignificant(r Mesodesmatidae 1 1 0.5 1 0.4306, 12 d.f.). Veneridae 5 3 0.375 2 We also comparedthe numberof species containingin- Crassatellidae 14 6 0.300 3 completedrillholes with the number of species successfully Limopsidae 17 7 0.292 3 drilledfor each formation.This ratioaveraged 0.26 forthe Kellielidae 200 71 0.262 2 Cretaceous and Paleocene (0.19 withoutthe Matthews Ungulinidae 3 1 0.25 1 Landing,which contained relatively few bivalve species) Carditidae 38 11 0.224 5 and 0.50 forthe Eocene and Oligocene(0.31 withoutthe Nuculanidae 21 6 0.222 4 Red Bluff,where ten specieshad incompletedrillholes but Corbulidae 688 167 0.195 24 onlysix were successfullydrilled). The patternis similar Cucullaeidae 12 2 0.143 2 whenonly infaunal bivalve prey are considered. Astartidae 42 5 0.106 3 Table 6 lists all bivalvetaxa that exhibitedincomplete Arcidae 36 4 0.1 2 drilling,along withdata on preyeffectiveness. Sixteen in- Mactridae 69 6 0.08 2 faunalfamilies of bivalves displayed incomplete drillholes Lucinidae 307 11 0.035 3 (Table 7). Prey effectivenessfor each of those families Noetidae 311 10 0.031 4 (excluding4 familieswith fewer than 10 total drillholes) Total for16 families 1764 312 0.150 62 rangedfrom 0.03 to 0.30. (These values are based onlyon Gastropods species that displayedincomplete drilling.) In particular, Strombidae 3 2 0.40 1 a large numberof corbulidspecies displayedincomplete Retusidae 5 1 0.167 1 drillholes.Prey effectivenessfor the Corbulidaewas 0.20 Naticidae 67 10 0.130 4 (forspecies that includedincomplete holes; whenall cor- Turritellidae 283 10 0.034 5 bulidspecies are included,prey effectiveness equals 0.167). Buccinidae 211 1 0.005 1 Otherfamilies with abundant drillholes and highprey ef- Total for5 families 569 24 0.040 12 fectivenessinclude the Carditidaeand Kelliellidae.

------

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 368 KELLEY & HANSEN

TABLE8-Numbers of complete(C.D.) and incomplete(inc.) drillholes and preyeffectiveness (P.E. = numberof incompleteholes: total attempted holes) for all bivalvesand forinfaunal bivalves within each level.

All bivalves Infaunalbivalves Level C.D. Inc. P.E. C.D. Inc. P.E.

Ripley 174 10 0.054 172 10 0.055 Providence 62 1 0.016 62 1 0.016 Corsicana 8 3 0.273 8 1 0.111 Kincaid 18 1 0.053 18 1 0.053 Brightseat 285 27 0.087 284 27 0.087 MatthewsLanding 13 4 0.235 13 4 0.235 Bells Landing 0 1 1.0 0 1 1.0 Bashi 233 11 0.045 233 11 0.045 Cook Mountain 721 48 0.062 668 39 0.055 Gosport 17 2 0.105 16 2 0.111 MoodysBranch 587 127 0.178 583 127 0.179 Red Bluff 258 60 0.189 258 55 0.176 MintSpring 163 25 0.133 158 25 0.137 Byram 46 8 0.148 46 8 0.148 Summary 2585 328 0.113 2519 312 0.110

TABLE9-Numbers and percentagesof all gastropodspecies, successfullydrilled gastropod species, infaunal gastropod species (inf.sp.), and infaunalsuccessfully drilled gastropod species (inf.dr. sp.) bearing incomplete and multipleholes.

All species Drilledsp. Inf.sp. Inf.dr. sp. # w/ # w/ % w/ % w/ % w/ % w/ % w/ % w/ % w/ % w/ Level inc. mul. inc. mul. inc. mul. inc. mul. inc. mul.

Ripley 1 2 3.4 6.9 6.7 13.3 0 0 0 0 Providence 0 0 0 0 0 0 0 0 0 0 Brightseat 0 0 0 0 0 0 0 0 0 0 MatthewsLanding 0 5 0 25.0 0 50.0 0 20.0 0 25.0 Bells Landing 1 1 3.8 3.8 10.0 10.0 7.1 7.1 14.3 14.3 Bashi 1 1 3.1 3.1 7.1 7.1 0 8.3 0 16.7 Cook Mountain 12 8 20.3 13.6 46.1 30.8 28.6 17.9 57.1 35.7 Gosport 1 1 4.3 4.3 20.0 20.0 10.0 10.0 20.0 20.0 MoodysBranch 2 4 4.3 8.6 10.5 21.1 7.7 7.7 12.5 12.5 Red Bluff 2 2 5.1 5.1 22.2 22.2 7.7 7.7 50.0 50.0 MintSpring 3 4 5.9 7.8 11.1 14.8 0 0 0 0 Byram 0 2 0 8.0 0 16.7 0 11.1 0 20.0

Besides consideringeffectiveness of individual prey taxa, but in the Oligoceneit was higherthan in the Cretaceous we computedthe effectivenessof all bivalve preywithin (x2= 19.42),Paleocene (x2= 9.21),and Eocene (x2= 13.51). each formation(Table 8). Prey effectivenesswas 0.11 for Spearman's rank correlationbetween stratigraphic order the entireCretaceous through Oligocene bivalve data set and preyeffectiveness (infaunal prey only) of a formation (both forinfaunal and all bivalves).Prey effectiveness for is statisticallysignificant (0.6168, P < 0.05). the Cretaceousand Paleocene was generallylow (less than A much lower incidence of incompletedrillholes oc- 0.1), withthe exceptionof levels that contained few drilled curredin gastropodsthan in bivalves(Table 9). Incomplete specimens.Prey effectivenessremained low in the Eocene drillholesare especiallyrare in Cretaceousand Paleocene Bashi and Cook MountainLevels (0.04-0.06),but was much samples; onlytwo of 105 gastropodspecies (4 incomplete higherfor upper Eocene and Oligocene samples (0.13- drillholes,total) examinedfrom these levels containedin- 0.19). In general,prey effectiveness in the Paleocene was completeholes. For mostEocene and Oligocenesamples, not significantlyhigher than in the Cretaceous(x2 = 3.37), about 3-6% ofthe species (0-10% ofthe infaunalspecies)

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 369

TABLE 10-Numbers of complete (C.D.) and incomplete(inc.) drillholesand preyeffectiveness (P.E. = numberof incompleteholes: total attemptedholes) forgastropod species that exhibitedincomplete drilling.

Level Species C.D. Inc. P.E.

Ripley Creonellatriplicata 3 1 0.25 Bells Landing Mesalia alabamiensistetradeiras 222 3 0.013 Bashi Eopleurotomacf. E. sayi 1 1 0.5 Cook Mountain Turritellacarinata 7 2 0.222 Turritellanasuta 21 2 0.087 Calyptraphorusvelatus 3 2 0.40 Hipponixpygmaeus 4 2 0.333 Calyptraeaaperta 2 1 0.333 Neveritalimula 11 7 0.389 Polinicesaratus 48 1 0.020 Naticid 3 1 0.25 Buccitritontexanum 211 1 0.005 Conimitratexana 0 2 1.0 Turrid 1 1 0.5 Retusa (Cylichna)galba 5 1 0.167 Gosport Mesalia vetusta 9 2 0.182 MoodysBranch Hipponixpygmaea 40 22 0.355 Euspira jacksonensis 5 1 0.167 Capulus americanus 2 3 0.6 Red Bluff Turritellaaff. T. premimetes 24 1 0.04 Tropisurculacaseyi 0 1? MintSpring Vexillumlintoidea 3 2 0.667 Syntomodrilliatantula 3 1 0.333 Pyramidella(Voluspa) leafensis 11 3 0.273 Byram Odostomiabyramensis 0 5 1.0

TABLE 1 1 -Numbers ofcomplete (C.D.) and incomplete TABLE 12-Numbers of attempteddrillholes (tot.), those (inc.) drillholesand preyeffectiveness (P.E. = numberof occurringin multiplybored valves (mult.) and the fre- incompleteholes: total attemptedholes) forall gastro- quency of multipledrillholes (M/T = numberof holes in pods and forinfaunal gastropods withineach level. multiplydrilled valves: total attemptedholes) for all bivalves and forinfaunal bivalves withineach level. All gastropods Infaunalgastropods Level C.D. Inc. P.E. C.D. Inc. P.E. All bivalves Infaunalbivalves Level Tot. Mult. M/T Tot. Mult. M/T Ripley 90 1 0.011 48 0 0 Providence 31 0 0 26 0 0 Ripley 184 0 0 182 0 0 Brightseat 156 0 0 133 0 0 Providence 63 0 0 63 0 0 MatthewsLanding 138 0 0 88 0 0 Corsicana 11 0 0 9 0 0 Bells Landing 261 3 0.011 253 3 0.012 Kincaid 19 2 0.105 19 2 0.105 Bashi 55 1 0.018 39 0 0 Brightseat 312 8 0.026 311 8 0.026 Cook Mountain 447 23 0.049 426 17 0.038 MatthewsLanding 17 2 0.118 17 2 0.118 Gosport 15 2 0.118 15 2 0.118 Bells Landing 1 0 0 1 0 0 MoodysBranch 104 36 0.257 48 1 0.020 Bashi 244 16 0.066 244 16 0.066 Red Bluff 39 1 0.025 23 1 0.042 Cook Mountain 769 30 0.039 707 16 0.023 MintSpring 107 6 0.053 61 0 0 Gosport 19 2 0.105 18 2 0.111 Byram 31 5 0.139 14 0 0 MoodysBranch 714 64 0.090 710 64 0.090 Summary 1474 78 0.050 1100 24 0.021 Red Bluff 318 49 0.154 313 45 0.144 MintSpring 188 29 0.154 183 29 0.158 Byram 54 4 0.074 54 4 0.074 Summary 2913 206 0.071 2831 188 0.066

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 370 KELLEY & HANSEN

TABLE 13-Numbers of attempteddrillholes (tot.), those occurringin multiply bored valves (mul.),and thefrequency of multipledrillholes (M/T = numberof holes in multiplydrilled valves: total attemptedholes) forbivalve species that exhibitedmultiply drilled valves.

Level Species Tot. Mul. M/T

Kincaid Striarcasp. 2 2 1.0 Brightseat Vokesulaaldrichi 35 4 0.114 Venericardiasp. 39 2 0.051 Phacoidessp. 95 2 0.021 MatthewsLanding Corbulasubcompressa 4 2 0.5 Bashi Venericardiahoratiana 134 2 0.015 Corbulasubengonata 39 8 0.205 Vokesulaaldrichi 51 6 0.118 Cook Mountain Pachecoa decisa 271 2 0.007 Pachecoa declivis 12 2 0.167 Cubitostreasellaeformis 60 14 0.233 Lucina pomilia 153 2 0.013 Corbulacf. C. alabamiensis 6 2 0.333 Corbulaextenuata 45 2 0.044 Caryocorbuladeusseni 38 6 0.158 Gosport cf.Caestocorbula murchisonii 3 2 0.667 MoodysBranch Caestocorbulawailesiana 82 17 0.207 Caryocorbuladensata 35 8 0.229 Venericardiadiversidensata 21 2 0.095 Alveinusminutus 225 9 0.04 Limopsisradiata 10 2 0.2 Lucina subcurta 104 4 0.038 Lucina curta 61 2 0.033 Bathytormusfiexurus 9 2 0.222 Spisula jacksonensis 41 2 0.049 Kelliella boettgeri 46 10 0.217 Verticordiacossmani 4 2 0.5 Nucula spheniopsis 18 4 0.222 Red Bluff Eburneopectensubminutus 5 4 0.80 Timothynusturgida? 4 4 1.0 Astartetriangulata 45 2 0.044 Corbula(V.) rufaripa 214 29 0.136 Corbula(Caryocorbula) engonata 12 10 0.833 MintSpring Corbula(Caryocorbula) engonata 19 2 0.105 Corbula(Varicorbula) laqueata 98 27 0.276 Byram Corbula(Varicorbula) laqueata 29 4 0.137

displayed incompletedrillholes. (The large Cook Moun- displayedincomplete holes, but onlythe Turritellidaeand tain sample is the anomaly,with 20% of all gastropod Naticidae had as manyas ten incompleteholes. Prey ef- speciesand 28% ofinfaunal species exhibiting incomplete fectivenesswas 0.03 and 0.13 forthe two families,respec- holes.) Whenthe numberof species with incomplete holes tively.(These values include only species containingin- is compared to that of successfullydrilled species, the completedrillholes; the value forall turritellidspecies is disparitybetween Cretaceous-Paleocene and Eocene-Oli- 0.013.) gocenelevels remains. However, there is a significantcor- The effectivenessof gastropodprey was muchless than relation between the percentof gastropodspecies with for bivalve prey (Table 11). Consideringall levels com- incompleteholes and number of gastropod specimens bined, only 78 of 1552 attempteddrillholes were incom- withina stratigraphiclevel (r = 0.7459, P < 0.01). The plete,yielding a preyeffectiveness of 0.05 (0.02 forinfaunal fewernumbers of species with incompletedrillholes in prey).Prey effectivenessin the Cretaceousand Paleocene older samples may be due in part to samplingeffects. was verylow (0 forall but the Ripley and Bells Landing, Table 10 lists gastropodspecies that exhibitedincom- whichhad a preyeffectiveness of 0.01). Prey effectiveness plete drillholes,along with data on preyeffectiveness; Ta- was higherin the Eocene and Oligocene,particularly in ble 7 providesa summaryby family.Five infaunalfamilies the Gosport,Moodys Branch,and ByramFormations. A

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 371

TABLE 14-Data fromTables 13 and 16, summarized TABLE 15-Numbers of attempteddrillholes (tot.), those by family.(Only familieswith infaunalspecies listed.) occurringin multiplybored shells (mult.) and the fre- Also shown, foreach family,is the numberof species quency of multipledrillholes (M/T = numberof holes in with multiplybored shells. (Species occurringat more multiplydrilled shells: total attemptedholes) forall gas- than one level are counted more than once.) Families tropods and forinfaunal gastropods withineach level. listed in order of frequencyof multipleholes. All gastropods Infaunalgastropods No. Level Tot. Mult. M/T Tot. Mult. M/T of spe- Ripley 91 4 0.044 48 0 0 Family Tot. Mul. M/T cies Providence 31 0 0 26 0 0 0 0 Bivalves Brightseat 156 0 0 133 MatthewsLanding 138 47 0.341 88 31 0.352 Ungulinidae 4 4 1 1 Bells Landing 264 4 0.015 256 4 0.016 Verticordiidae 4 2 0.5 1 Bashi 56 2 0.036 39 2 0.051 Nuculidae 18 4 0.222 1 Cook Mountain 470 48 0.102 443 39 0.088 Crassatellidae 9 2 0.222 1 Gosport 17 2 0.118 17 2 0.118 Limopsidae 10 2 0.2 1 MoodysBranch 140 25 0.179 49 2 0.041 Corbulidae 710 129 0.182 15 Red Bluff 40 9 0.225 24 5 0.208 Kelliellidae 271 19 0.070 2 MintSpring 113 18 0.159 61 0 0 Mactridae 41 2 0.049 1 Byram 36 7 0.194 14 2 0.143 Astartidae 45 2 0.044 1 Carditidae 194 6 0.031 3 Summary 1552 166 0.107 1124 89 0.079 Lucinidae 413 10 0.024 4 Noetidae 285 6 0.021 3 Total for12 families 2004 188 0.094 34 bored valves forthe Cretaceous,Paleocene, Eocene, and Gastropods Oligocene,respectively. Chi-squared tests indicate that the value foreach of these is than Strombidae 5 3 0.60 1 epochs significantlygreater that for the precedingepoch, with the overall increase Terebridae 5 2 0.40 1 = Vermetidae 14 2 0.142 1 beinghighly significant (x2 36.26, 1 d.f.).The patternis Turritellidae 543 76 0.140 9 the same if onlyinfaunal bivalves are considered. Buccinidae 212 6 0.028 1 Table 13 liststhe bivalvespecies that included multiply bored shells. Twelve infaunalbivalve familiesare Total for5 families 779 89 0.114 13 represented(Table 14),though the Ungulinidae, Verticordiidae, Crassatellidae,and Limopsidae are representedby ten or fewerdrillholes each. The incidenceof multipleholes for significantrank correlation(0.8531, P < 0.01) exists be- the othereight families ranged from about 2-4 %, withthe tweenstratigraphic order of levels and preyeffectiveness, exceptionof the Corbulidaeand Nuculidae. Onlyone nu- suggestinga temporalincrease in preyeffectiveness. culid species exhibitedmultiple boreholes (4 of 18 bore- Multiple drillholeswere less commonthan incomplete holeswere in multiplybored valves). However,15 corbulid holes,but wereprominent for certain levels and taxa (Ta- species displayedmultiple boreholes; 129 drillholes(18% ) ble 5). No multiplybored preywere noted in any of the occurredin multiplybored shells (includingspecies with Cretaceoussamples of bivalves.With the exceptionof the onlysingly bored valves,the value is still 15%). Bells Landing,which contains relatively few bivalves, mul- In general,fewer gastropod than bivalvespecimens ex- tiplybored species were present in all thePaleocene through hibitedmultiple drillholes. Multiply drilled gastropod spe- Oligocenesamples. The percentof species withmultiple cies were most commonin the MatthewsLanding, Cook boreholesranged from 2-23%, withhigh percentages oc- Mountain,Moodys Branch,and Oligocenelevels (Table curringin the MatthewsLanding, Bashi, MoodysBranch, 9). The occurrenceof multiplybored species is not simply and Red Blufflevels. The incidence of multiplybored relatedto sampling,as the correlationbetween sample size species may be related to sampling,for the correlation and numberof species with multiplybored shells at a betweenpercent of species with multiple holes and number stratigraphiclevel is nonsignificant(r = 0.3019, 12 d.f.). ofspecimens in a unitis significant(r = 0.5425,P < 0.05). There is some evidencefor a temporalincrease in the Consideringall fourteenlevels, about 7 % ofall drillholes incidenceof multiple drillholes in gastropods. For the twelve occurredin multiplybored valves (Table 12). A significant levels fromwhich gastropod data were available, 11% of rank correlationoccurs betweenstratigraphic order of a all drillholesoccurred in multiplybored shells (Table 15). level and the percentof drillholesoccurring in multiply Only 3% of Cretaceous drillholesoccurred in multiply bored valves (0.6964, P < 0.01). Multiple drillholesthus bored shells; the values for the Paleocene, Eocene, and appear to have becomemore common through time, with Oligoceneare 9%, 11%, and 18%, respectively.The dif- 0, 3%, 6%, and 15% of drillholesoccurring in multiply ferencesbetween Cretaceous and Paleocene levels (x2 =

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 372 KELLEY & HANSEN

TABLE 16-Numbers of attempteddrillholes (tot.), those occurringin multiply bored shells (mul.),and the frequency of multipledrillholes (M/T = numberof holes in multiplydrilled shells: total attemptedholes) forgastropod species that exhibitedmultiply drilled shells.

Level Species Tot. Mul. M/T

Ripley Laxispira monilifera 14 2 0.143 Creonellatriplicata 4 2 0.5 MatthewsLanding Rissoinaalabamensis 30 10 0.333 Turritellaaldrichi 93 31 0.333 Fasciolaridae 6 2 0.333 Cancellaridae 6 2 0.333 Microdrilliasp. 11 2 0.182 Bells Landing Mesalia alabamiensistetradeiras 225 4 0.018 Bashi Turritellagilberti 28 2 0.071 Cook Mountain Turritellacarinata 9 6 0.667 Turritellanasuta 23 4 0.174 Turritellasp. 103 20 0.194 Calyptraphorusvelatus 5 3 0.6 Hipponixpygmaeus 6 5 0.833 Buccitritontexanum 212 6 0.028 Eucheilodonreticularis 6 2 0.333 Hesperiturrisnodocarinatus 3 2 0.667 Gosport Mesalia vetusta 11 2 0.182 MoodysBranch Pyramidellameyeri 2 2 1.0 Turritellaalveata 26 2 0.077 Hipponixpygmaea 72 18 0.25 Capulus americanus 5 3 0.6 Red Bluff Turritellaaff. T. premimetes 25 5 0.20 Syntomodrilliacolorubra 7 4 0.571 MintSpring Tritiariafalsus 3 3 1.0 Vexillumlintoidea 5 3 0.6 Pyramidella(Voluspa) leafensis 14 7 0.5 Eulimellaclearyensis 5 5 1.0 Byram Terebrasp. 5 2 0.4 Odostomiabyramensis 5 5 1.0

4.08) and Eocene and Oligocenelevels (x2 = 4.50) are sta- As expected,based on the life mode of the predator, tisticallysignificant. Spearman's rank correlation between infaunalmolluscs are the mostcommonly drilled prey; of stratigraphicorder of a level and percentof drillholesoc- thisgroup, some species are drilledin substantiallygreater curringin multiplybored shells also is significant(0.6276, proportionsthan expectedbased on theirrepresentation P < 0.05). in the fauna. Individualsof the familyTurritellidae, for Table 16 lists the gastropodspecies that experienced example,are particularlyabundant in Paleocene assem- multipleboring. Infaunal species representfive different blages. In the BrightseatFormation, Haustator gnoma families(Table 14). Multiple drillholeswere particularly comprises21% of the fauna and 32% of the drilledindi- commonin the Turritellidae.Of the 543 drillholeson tur- viduals.In the MatthewsLanding Member, Turritella al- ritellidspecies withmultiple drillholes, 14% werein mul- drichi(28% ofthe fauna,44% ofdrilled individuals) and, tiplybored shells. (When species with only singlybored in theBells LandingMember, Mesalia alabamiensistetra- shells are included,the value is 6%.) deiras (45% of fauna, 67% of drilled) conformto this pattern. HEAVILY PREYED TAXA THROUGH TIME Other commonlyattacked familiesare the Lucinidae and Corbulidae.Species of these familiesare oftendrilled Extantnaticids are selectivewith respect to preytaxon, at greaterthan averagefrequencies, whether or not they as weretheir counterparts studied from various, primarily are the mostabundant taxa. All threefamilies are infaunal Neogene,fossil assemblages (Kitchell, 1982; Hoffmanand suspension feeders and sluggish burrowers.Other fre- Martinell,1984; Kitchellet al., 1981; Kelley, 1988, 1991). quentlydrilled taxa are the Arcidae,Noetiidae, and Car- Taxon-selectivepredation also is evidentin ourCretaceous ditidae,and the Crassatellacea.All are nonsiphonate,in- throughOligocene data set. faunal suspension feedingbivalves, and extremelyslow

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 373

burrowers.Certain species of otherfamilies, such as the pothesis of escalation. One predictionmay be that the Veneridaeor Buccinidae,may be drilledsubstantially in frequencyof drilling should have increasedduring the Ce- a particularenvironment within a formation,but not as nozoicas a resultof increasing biological hazards (Vermeij, consistentlythrough time as the groupsmentioned above. 1987). Anotherinterpretation suggests that, if escalation Infaunalspecies representedby morethan 100 individ- involvingprey response to predatorsoccurred, one might uals in any formationbut withlower than averagedrilling expectan initialincrease in drillingfrequency followed by frequenciesinclude Nucula, Hilgardia, Yoldia (all depos- stabilizationof frequenciesafter parity was reached.This it-feedingbivalves), Limopsis (nonsiphonatesuspension explanationwas mentionedby Allmonet al. (1990), who feedingbivalve), Spisula, Alveinus,and Kelliella (all three suggestedthat turritellineantipredatory armor evolved siphonatesuspension feeding bivalves). Sander and Lalli quicklyin the Cretaceousand remainedrelatively static (1982) noted the conspicuouslack of drillholesin Nucu- throughthe Cenozoic. Our data on overall drillingfre- lana and Limopsis in modernmolluscan assemblages from quencies showneither of these patterns.For the 30+ mil- Barbados, citingthe small size of these two taxa as a pos- lionyear time span fromthe Late Cretaceousinto the early sibledeterrent. In ourdata set thesegenera are comparable Oligocene,we see neithera progressiveincrease nor sta- in size to heavilydrilled taxa, so anotherexplanation seems bilization;we see statisticallysignificant fluctuation. necessary.Limopsis resemblesmany lucinids in size, but Vermeij (1987) also suggestedthat escalation may be Limopsis has a thickershell, probably increasing its cost- episodic;it shouldoccur during periods of warming, trans- benefitratio. It also has a hairyperiostracum, which may gression,and increasedproductivity, and be interrupted help protectit (Vermeij,pers. comm.). Nuculana (and by extinctionepisodes. The stratigraphicdistribution of most membersof the Nuculanidae) have conspicuously our samples is still too spottyto correlatereliably with thin shells,but theyare active burrowersand thus more this complexof environmentalvariables. It is intriguing, likelyto avoid capture.Alveinus and Kelliella, however, however,that the drop in predationfrequency in the late are verysmall. Eocene coincided with a diversitydecrease for molluscs We have notyet examined whether the dynamics of prey (Hansen, 1988) and an episode of generalclimatic cooling selectivity,as recordedin our database, changedthrough (Keller, 1983; Wolfeand Poore, 1982). There are also sig- time.Ongoing cost-benefit analyses will indicatewhether nificantincreases in drillingfrequencies soon after ex- the degreeof preyselectivity, or the criteriaused in prey tinctionepisodes at the Cretaceous-Tertiaryand Eocene- selection,changed duringthe Cretaceousthrough Oligo- Oligoceneboundaries. We are in the processof obtaining cene. additionalsamples to definebetter the recordof naticid predationin the contextof these environmentalchanges. DISCUSSION Escalationmay predict a temporalincrease in stereotypy of naticid behavior,because stereotypyin general pro- Our data showa dramaticand abruptincrease in overall motespredator efficiency (Kelley, 1988). Thus farwe have drillingfrequency in the middleDanian (lowerPaleocene) examinedonly stereotypy with respect to selectionof drill- BrightseatFormation. We cannotbe sure if the low inci- hole site; futurework will also examineselectivity of prey dence of drillingin the earlyDanian Kincaid Formation size and taxon, using cost-benefitanalysis. There is no is real or an artifactof preservationand verylow sample significantincrease in stereotypyof drillhole site with time; size. In any case, our data suggestthat naticid drilling naticidsfrom all stratigraphiclevels were selective of drill- frequenciesrose relatively abruptly and soon afterthe Cre- hole site formost preytaxa. Lack of drillholesite selec- taceous-Tertiaryextinction. Hazel et al. (1984) deter- tivitywas commononly in the Corbulidae.Although An- mined that the calcareousnannofossil assemblage of the derson(1992) has suggestedthat corbulids are not unique BrightseatFormation is representativeof the Chiasmo- with regardto patternsof drillholesite selectivity,the lithus danicus zone (NP3, middle Danian stage). Using presentstudy supports Kelley's (1988) contentionthat cor- the Graphic CorrelationTechnique, Hazel et al. (1984) bulids are oftenanomalous in termsof drillholesite and calculated that the durationof the NP3 zone was from preyeffectiveness. about 65.4 Ma to about 63.8 Ma. This would place the Escalation of preyeffectiveness is suggestedby signifi- dramaticincrease in drillingfrequencies to withinone to cant increases fromthe Cretaceous to the Oligocene in threemillion years of the Cretaceous-Tertiaryboundary. percentagesof incompleteand multipleboreholes in the This highfrequency of drilling was maintaineduntil the wholefauna. More directinformation may be providedby late Eocene, when it dropped to Cretaceous levels, and groups that are particularlyprone to naticid predation. then rose again in the Oligoceneto a moderatelevel. The The modelof escalation may predict that a taxonthat was Moodys Branch Formation,characterized by a verylow heavilyand consistentlypreyed upon fortens of millions drillingfrequency, represents the largest sample size (14,310 of yearswould have respondedeither with escalation, re- specimens),so this data point is robust.Apparently the strictionto refugia,or eventualextinction. The Turritel- breakthroughthat enabled the naticids to achieve high lidae, Lucinidae,and Corbulidaeusually suffered heavier drillingfrequencies in thePaleocene onlyraised the ceiling than averagepredation when present in an assemblagein on those predationintensities. substantialnumbers. The Turritellidaeand Corbulidae As suggestedby Kitchell(1990) and by Vermeij(1987), show significantincreases in percentagesof incomplete it is difficultto generatesimple predictionsfrom the hy- and multipleboreholes from the Cretaceousthrough the

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions 374 KELLEY & HANSEN

EarlyTertiary, but the increasesare notconsistent through BERG,C.J., and NISHENKO,S., 1975,Stereotypy of predatoryboring time.The greatestincreases for the Corbulidaeoccur from behaviorof Pleistocene naticid gastropods: Paleobiology, v. 1, p. theCretaceous to thePaleocene (coincidentwith the great- 258-260. BOGGS,C.H., RICE, J.A.,KITCHELL, J.A., and POST, W.M., 1984, Pre- est increasein predationrate) and are less dramaticin the dationat a snail'space: What'stime to a gastropod?:Oecologia, Eocene. The greatestincreases for the Turritellidaeoccur v. 62, p. 13-17. betweenthe Paleocene and Eocene. The Lucinidae do not BRETSKY,S.S., 1976,Evolution and Classificationof the Lucinidae show increasesin preyeffectiveness; they always have ex- (Mollusca;Bivalvia): Palaeontographica Americana, v. 8, no. 50, tremelysmall numbersof incompleteand multipledrills. 337 p. and by naticacean Much workremains to be done on these predation-prone CARRIKER,M.R., 1981,Shell penetration feeding and muricaceanpredatory gastropods: A synthesis:Malacologia, groups,however, as yetwe do not have cost-benefitanal- v. 20, p. 403-422. ysesof their component species nor do we understandtheir DEANGELIS, D.L., KITCHELL,J.A., and POST, W.M., in press, The phylogeneticrelationships. natureof stasisand change:Potential coevolutionary dynamics, in STENSETH,N., ed.,Coevolution of Ecosystems. Cambridge Uni- versityPress, Cambridge. CONCLUSIONS DOCKERY,D.T., III, 1977,Mollusca of the Moodys Branch Formation, Mississippi:Mississippi Geological Survey Bulletin, v. 120,212 p. 1. Frequencyof drillingby naticidgastropods shows nei- DOCKERY,D.T., III, 1982,Lower Oligocene Bivalvia of the Vicksburg thera progressiveincrease nor increase followed by sta- Groupin Mississippi:Mississippi Bureau of GeologyBulletin, v. bilization,as has been predictedbased on the hypoth- 123,261 p. esis of escalation. Drillingfrequencies were relatively DUDLEY, E.C., and VERMEIJ,G.J., 1978, Predation in timeand space: low in the Cretaceous,declined furtherat the Creta- Drillingin the gastropodTurritella: Paleobiology, v. 4, p. 436- rose above the 441. ceous-Tertiaryboundary, dramatically HANSEN,T.A., 1988,Early Tertiary radiation of marine molluscs and boundary,and remainedhigh until the late Eocene. the long-termeffects of the Cretaceous-Tertiaryextinction: Pa- Drillingdeclined significantly at the Eocene-Oligocene leobiology,v. 14,p. 37-51. boundaryand thenincreased during the Oligocene,but HAZEL, J.E., EDWARDS,L.E., and BYBELL, L.M., 1984, Significant did not reachthe highlevels of the Paleocene and early unconformitiesand thehiatuses represented by themin thePa- leogeneof the Atlantic and Gulf Coastal Province, in SCHLEE,J.S., Eocene. Accumula- to based on the of ed., InterregionalUnconformities and Hydrocarbon 2. Contrary our predictions hypothesis tion:American Association of PetroleumGeologists Memoir 36, no trendsoccurred in behavioral escalation, temporal p. 59-66. stereotypyof naticids,as evidencedby drillholesite. HOFFMAN,A., and MARTINELL,J., 1984, Prey selection by naticid 3. Frequencyof incompletedrillholes and multiplybored gastropodsin thePliocene of Emporda (Northeast Spain): Neues shells increasedboth withinentire faunas and individ- Jahrbuchfur Geologie and PalaontologieMonatshefte, v. 1984,p. ual an increasein preyeffectiveness 393-399. lineages,indicating Pleistocene consistentwith the of escalation. KABAT,A.R., and KOHN,A.J., 1986, Predation on early hypothesis in Palaeogeography, Palaeoclimatology, such as and naticid gastropods Fiji: 4. Certain prey taxa, turritellids,corbulids, Palaeoecology,v. 53, p. 255-269. lucinids,were consistentlypreferred by naticids.Pre- KELLER,G., 1983,Biochronology and paleoclimaticimplications of ferredprey generally were sluggishburrowers, though Middle Eocene to Oligocene plankticforaminiferal faunas: Marine not all such taxa wereheavily drilled. Micropaleontology,v. 7, p. 463-486. KELLEY, P.H., 1988,Predation by Miocenegastropods of the Ches- apeake Group:Stereotyped and predictable:PALAIOS, v. 3, p. ACKNOWLEDGMENTS 436-448. KELLEY, P.H., 1989, Evolutionarytrends within bivalve preyof Ches- We thankDavid Dockeryfor help in locatingcollecting apeakeGroup naticid gastropods: Historical Biology, v. 2, p. 139- sites and David Haasl for assistance in the field.David 156. Haasl, Ben Farrell,Elizabeth Akins, and Ralph Lattimore KELLEY, P.H., 1991,Cannibalism by Chesapeake Group naticid gas- assisted in data tabulation and analysis. Norman Sohl, tropods: A predictable resultof stereotypedpredation: Journal of 75-80. Warren and Lauck Ward providedaccess Paleontology,v. 65, p. Blow, graciously KITCHELL,J.A., 1982, Coevolution in a predator-preysystem: Pro- to additional samples. The manuscriptbenefitted from ceedings of the North American Paleontological Convention III, reviewsby GeeratVermeij, Warren Allmon, and an anon- v. 2, p. 301-305. ymousreviewer. Support forthis studywas providedby KITCHELL,J.A., 1986, The evolutionof predator-prey behavior: Na- National Science Foundationgrant EAR 8915725. ticid gastropods and their molluscan prey, in NITECKI,M., and KITCHELL, J.A.,eds., Evolution of Animal Behavior: Paleontolog- ical and Field Approaches: Oxford UniversityPress, Oxford,p. REFERENCES 88-110. KITCHELL,J.A., 1990, The reciprocal interaction of organism and ADEGOKE,O.S., and TEVESZ,M.J.S., 1974, Gastropod predation pat- effectiveenvironment: Learning more about "and," in Ross,R.M., terns in the Eocene of Nigeria: Lethaia, v. 7, p. 17-24. and ALLMON,W.D., eds.,Causes of Evolution:A Paleontological ALLMON,W.D., NIEH, J.C., and NORRIS,R.D., 1990, Drillingand peel- Perspective:University of Chicago Press, Chicago, p. 151-169. ingof turritelline gastropods since the Late Cretaceous:Palaeon- KITCHELL,J.A., BOGGS,C.H., KITCHELL,J.F., and RICE, J.A., 1981, tology,v. 33, p. 595-611. Prey selection by naticid gastropods:Experimental tests and ap- ANDERSON,L., 1992,Naticid gastropodpredation on corbulidbi- plication to the fossil record: Paleobiology, v. 7, p. 533-552. valves:Effects of physical factors, morphologic features, and sta- KITCHELL,J.A., BOGGS,C.H., RICE, J.A., KITCHELL,J.F., HOFFMAN, tisticalartifacts: PALAIOS, v. 7, p. 602-620. A., and MARTINELL,J., 1986, Anomalies in naticid predatorybe-

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions NATICIDGASTROPOD ESCALATION 375

havior:A critiqueand experimentalobservations: Malacologia, v. VANVALEN, L., 1973,A newevolutionary law: Evolutionary Theory, 27,p. 291-298. v. 1, p. 1-30. MACNEIL,F.S., and DOCKERY, D.T., III, 1984, Lower Oligocene Gas- VERMEIJ,G.J., 1983, Intimate associations and coevolutionin the sea, tropoda,Scaphopoda, and Cephalopodaof the VicksburgGroup in FUTUYMA,D.J., and SLATKIN,M., eds., Coevolution: Sinauer in Mississippi:Mississippi Bureau of GeologyBulletin, v. 124, Associates, Sunderland, Massachusetts, p. 311-327. 416 p. VERMEIJ,G.J., 1987, Evolution and Escalation: An Ecological History PALMER,K.V.W., and BRANN, D.C., 1965, 1966, Catalogue of the of Life: Princeton UniversityPress, Princeton,527 p. Paleoceneand Eocene Mollusca of the Southern and Eastern Unit- VERMEIJ, G.J., DUDLEY, E.C., and ZIPSER,E., 1989, Successful and ed States:Bulletin of American Paleontology 48, v. 218,1027 p. unsuccessfuldrilling predation in Recentpelecypods: Veliger, v. Ross, R.W.,and ALLMON,W.D., eds., 1990,Causes of Evolution:A 32, p. 266-273. PaleontologicalPerspective: University of Chicago Press, Chicago, VERMEIJ, G.J.,ZIPSER, E., and DUDLEY,E.C., 1980, Predation in time 479 p. and space:Peeling and drillingin terebridgastropods. Paleobiol- SANDER,F., and LALLI,C.M., 1982, A comparativestudy of the mol- ogy,v. 6, p. 352-364. lusk communitieson the shelf-slopemargin of Barbados,West WOLFE,J.A., and POORE, R.Z., 1982, Tertiarymarine and nonmarine Indies:Veliger, v. 24,p. 309-318. climatic trends,in BERGER, W.H., and CROWELL, J.C., eds., Cli- SOHL,N.F., 1969,The fossilrecord of shell boring by snails: American matein EarthHistory: National Academy of Sciences, Washing- Zoologist,v. 9, p. 725-734. ton,D.C., p. 154-158. STEPHENSON,L.W., 1941, The larger invertebratefossils of the Na- varro Group of Texas: Universityof Texas, Publication No. 4101, ACCEPTED AUGUST 13, 1992 641 p.

.11

This content downloaded from 140.160.178.168 on Thu, 15 May 2014 18:17:29 PM All use subject to JSTOR Terms and Conditions