sign in sign up
<<
Home , Bottom feeder, Catastrophism, Morone, Paleoecology, Paleolimnology, Predatory dinoflagellate

Palaeogeography, Palaeoclimatology, Palaeoecology, 62 (1988): 577 592 577 Elsevier Science Publishers B.V., Amsterdam -- Printed in The Netherlands

FISH AND ENVIRONMENTAL INFERENCE IN

R. L. ELDER and G. R. SMITH

Department of , Oberlin College, Oberlin, OH 44074 (U.S.A.) Museum of , University of Michigan, Ann Arbor, MI 48109 (U.S.A.)

(Accepted for publication April 24, 1986)

Abstract

Elder, R. L. and Smith, G. R., 1988. taphonomy and environmental inference in paleolimnology. Palaeogeogr., Palaeoclimatol., Palaeoecol., 62: 577-592.

The contribution of fish studies to generally takes the form of (1) inference from analogies in modern fish faunas and (2) fish taphonomy -- the pattern of and dispersal of bones. (1) Modern fish faunas and associated provide taxonomic, ecological, or functional analogues for interpretation of ancient limiting factors and behaviors. These inferences presume taxonomic conservatism. They also presume functional relationships between morphological form and feeding mode or . They become weak with increased geologic age or phyletic distance between ancient subject and modern analogue. (2) Fish taphonomy may contribute information about , community composition, history, mortality, depositional environment, and preservation. Taphonomic reconstruc- tion of and preservation depends on the applicability of analogous processes in modern ecology and limnology. In aquatic taphonomy, temperature is the most important factor in determining the fate of a carcass. Above about 16°C (depending on depth and pressure), most carcasses are made buoyant by bacterial decay gases and are transported to the surface where they may decay further and fall piecemeal into deepwater environments, or drift to beach environments where wave energy disarticulates, abrades, and scatters the bones. Below about 16°C, most carcasses remain on the bottom until buried; they may be disturbed by , depending on concentration in the hypolimnion.

Introduction loss relative to the original, living community, and (2) discover critical similarities and differ- Paleoecology attempts to reconstruct the ences between modern analogues and the relationships of ancient organisms to their ancient organisms and environments. biological and physical environments. It draws There are two ways in which paleoecological its information from many different fields of reconstruction can be approached: (1) by study in order to make inferences about the inference from taxonomically or structurally ancient geologic setting, climate, community similar modern organisms used as analogues, structure, population dynamics, trophic re- which suggest limiting factors, behaviors, and lationships, and behaviors. Drawing reason- trophic dynamics; and (2) by taphonomy (espe- able inferences requires that we be able to (1) cially -- the events from death estimate the effects of physical and biological to burial): the limnological conditions, biologi- processes on the assemblage and sedi- cal processes, and depositional systems affect- ments, including evaluation of information ing life, death, and preservation. Each of these

0031-0182/88/$03.50 © 1988 Elsevier Science Publishers B.V. 578 approaches has advantages and disadvantages, Inferences based on size and shape of jaws and depending on the particular study. The infer- pharyngeal arches, and size and shape of teeth ences provided by one may sometimes test (canine, crushing, grinding, etc.) are often those of the other, accepted as reasonable (Grande, 1980). Following the precedent successful in paleo- botany (e.g. Dorf, 1930, 1970; Chaney et al., The use of modern analogues 1944; but see also Wolfe, 1971) it is assumed that if all living members of a or family We first consider the use of analogy, its are limited to a specific tolerance range that limitations, and its development in the fish fossil members of those groups probably shared paleoecological literature. Ecological charac- the same limitation. Confidence in the infer- teristics of modern species (analogues) are ence is a function of the number of living attributed to fossil forms to provide explana- species in the group, the sharpness of the tions. The analogues may be phyletic relatives limitation, and the absence of exceptions. For or functional counterparts. The most com- example, the primary families monly applied attributes are environmental (Myers, 1949) are thought to be indicators of a tolerances (, temperature, substrate, freshwater depositional environment. In a current or wave energy, turbidity, depth), mixed fossil assemblage, reliability of the behaviors (feeding, spawning), and trophic inference, however, depends on the proportion relationships. The reasonableness of the in- of primary freshwater representatives. ferred attribute depends on our confidence in Similarly, northern hemisphere families either the phylogenetic or functional related- whose species show rather sharp northern ness of the modern and fossil species. For limits that are correlated with summer iso- example, to infer past climates from Recent therms (Radforth, 1944) may be reliable indica- distributions, we must assume that tors of length of growing season or of lower of physiological tolerances in the past has been summer temperatures. Northern families at least as conservative as appears in Recent whose southern limits are correlated with representatives of the group in question, temperature may provide evidence bearing on Presently there are serious limitations in the maximum summer temperatures, provided that use of morphology to infer ecological param- moisture and elevation are considered (Cross, eter states. Most environmental parameters do 1970; Smith, 1978). Most such inferences have not affect the size and shape of with been based on fossil occurrences outside the sufficient regularity to allow confident infer- species' modern range. Integration of diverse ence of any conditions, for example, salinity or faunal evidence to reconstruct the seasonality temperature, from body shape. The relation- of a paleoenvironment is possible for Pleisto- ship between developmental temperature and cene faunas containing both northern and number of serial parts is partly understood southern Recent forms (Hibbard, 1955, 1960). (e.g. Barlow, 1961; Johnson and Barnett, 1975; faunas (Smith, C. L., 1962) and Lindsey and Arnason, 1981) but, because the faunas (Uyeno, 1963; Semken, 1966; means and the shapes of the curves vary Neff, 1975; Teller and Bardack, 1975; Teller- between species, no inference about fossil Marshall and Bardack, 1978) have also been species is possible with current knowledge. As studied in this way. a result of these limitations, inferences about Cross (1970) discussed limitations to using environment have been based primarily on fish alone to infer climate and showed that phyletic relatedness, many papers (e.g. Smith, C. L., 1954; Smith, The relationships of morphology to feeding G.R., 1963) that inferred cooler climates could behaviors and other aspects of predator-prey be interpreted otherwise to show wetter cli- relationships are somewhat better understood, mates. He showed that fish alone do not 579 allow us to distinguish between more water Temperature and climate and cooler water. Climatic interpretation of cooler climates in the past is strengthened, Climate and water temperatures were stud- however, when concordant with other ele- ied by Smith, C. L., (1954, 1964) and Smith, ments in the flora and fauna. As usual, these G.R. (1963) in investigations of Great Plains methods depend on degree of relationship, fish faunas relative to Late Cenozoic climatic number of recent species on which the gener- changes in North America. Interpretations of alization is based, and sample size in the fossil climates cooler than at present were based on assemblage, fossil occurrence of genera and species that are The reliability of the analogy approach to now cool stenothermic, such as perch (Perca), paleoecology depends on the accuracy of the muskellunge (Esox), and other northern underlying and functional mor- groups. Similarly, the sequence of Pleistocene phology. Inferences become less certain when faunas from Hansen Bluff, near Alamosa, older periods of geologic time are involved Colorado shows a local transition from and when large amounts of evolution have coolwater fish (Salmo, presently northern) to taken place. Because morphology is the result warmwater fish (Gila, presently southern) that of long-term responses to physiological and indicates a warming trend (Rogers et al., 1985). pressures, the method of Pleistocene climate on the northern plains was using analogies is not easily amenable to inferred by Sherrod (1963) on the basis of an direct experimentation or observation today; assemblage of fish bones in an ice-contact . the inferences are not easily falsifiable, but Cooler Pleistocene climates were inferred by are weakened by discovery of conflicting Swift (1968), Smith, M. L. (1981), and Smith, examples. M.L. et al. (1975) for various localities farther Although most papers on fossil fishes con- south in North America. Earlier literature was tain comments about habit or habitat, only a reviewed by Miller (1965). few authors have analyzed whole faunas or offered discussions of methods. Most applica- Stream gradient and habitat tions of analogy have been restricted to Late Cenozoic (David, 1944; Smith, C. L., 1954, 1964; Smith, G. R. and Lundberg (1972) attempted Smith, G. R., 1963, 1975; Gaudant, 1979a, b, to draw conclusions about stream gradient and 1981 and references therein), Early Cenozoic related factors affecting the depositional en- (Wilson, 1977, 1980; Grande, 1980; Buchheim vironment of the Sand Draw local fauna in pre- and Surdam, 1981) and Mesozoic (Schaeffer, glacial Nebraska. The data were drawn from 1970; Waldman, 1971; Sch~fer, 1972)faunas. both phyletic and structural analogies. The fish remains in the fauna belong to genera Salinity known from low-gradient, low-elevation habi- tats today. Furthermore, most of these genera, Ictiobus (buffalo fish), Hiodon (goldeye), Chae- The most prolific use of the analogy ap- nobryttus (warmouth sunfish), Lepomis proach to infer salinity in non-obvious circum- (sunfish), and Morone (white ), are deep- stances has been by Gaudant (1979a, b, 1981). bodied, laterally compressed fishes that grow He analyzed paleosalinity in a number of Early to moderate or large size. These shape and size Cenozoic European faunas and added infer- characteristics are found in fishes that inhabit ences about temperature and depth. The vari- slow-moving (low-gradient) waters usually at ous proportions of freshwater and marine taxa low elevations and usually large in volume. at different localities usually make it possible They are not common in high-gradient, small to determine the salinity of a depositional creeks (Smith, G. R. and Lundberg, 1972). environment. Gradient, elevation, habitat volume, deposi- 580

~~ ~ by Perkins (1970a) and especially by Grande (1980): flat, crushing teeth indicate mollusk and decapod feeding (Heliobatis); modern planktivory by paddlefish indicates the same in Crossopholis; predaceous teeth indicate preda- A",z~ .~.,., .,, ,,. 9~lll~,< ceous habits in Lepisosteus, Phareodus, Mio- plosus, and one species of Amia, while thick, short, styliform teeth indicate mollusk and Fig.1. Classification, by shape, of fishes from five low- gradient stream (A, in Michigan, Iowa, Kansas, crustacean feeding in another species of Amia; Oklahoma, Alabama; N=4235) and five high-gradient large pointed jaw teeth indicate on habitats (B, in British Columbia, Oregon, Idaho, Utah, small fish and in Eohiodon; planktivory New ; N= 2305). Low-gradient streams have high is inferred from clupeomorph mouth and body diversity of taxa and body shapes, with an abundance of deep-bodied fishes. High-gradient streams have low form (Knightia and young Diplomystus); ven- diversity of taxa and shapes, and few if any deep-bodied tral jaws and tooth loss indicate herbivory or inhabitants. Data from Division of Fishes, University of scavenging (Notogoneus); slightly coniform Michigan Museum of Zoology. pharyngeals in suckers indicate possible om- nivory (Amyzon); similar teeth indicate similar tional environment, and species richness are diet in fossil and modern , and in fossil related in complex ways of major importance and modern troutperch; small teeth indicate to interpretation of paleoenvironments (Fig. 1). small prey (Asineops); obtuse, grinding pharyn- Elevation, high gradient, and suspended load geal teeth indicate feeding on mollusks and (degradation), in concert, are generally nega- insects (Priscacara). Grande (1980) extended tively correlated with species richness, while these inferences to trophic generalizations, low gradient, increased habitat volume, and and Perkins (1970a) used his classification to stability are associated with high species draw conclusions about community equit- richness (Smith, G. R., 1981). Other interpreta- ability. tions of habitat volume or gradient based on Similarly, Smith, G. R. (1975) assigned tro- fishes include Smith, C. L. (1958), Uyeno and phic roles to Lake Idaho fishes based on Miller (1962), Schultz (1965), Hibbard and feeding morphology and modern analogues: Dalquest (1966), Lundberg (1967), Wilson large, carnivorous teeth and large mouth and (1968), Bell (1973), Ossian (1973), Eshelman body size indicate a top (Rhabdo- (1975), and Bennett (1979). fario); in suckers, a terminal mouth indicates a Another example of the use of the analogy midwater (Chasmistes) and a ven- approach to infer habitat is Buchheim and tral mouth indicates a benthic, small-particle Surdam's (1981) analysis of Lake Gosiute, in (Catostomus); large size, large jaws, which they infer paleoenvironments, based in and long canine teeth indicate a top carnivore part on modern analogies of the fish fauna. (Ptychocheilus); sharp, grinding pharyngeal This analysis is discussed in the taphonomy teeth indicate herbivory (Acrocheilus and section below. Orthodon); blunt, molariform teeth indicate molluskivores (Mylocheilus); small and Trophic relationships medium-sized, midwater carnivory is inferred from analogy to similar Recent relatives (Gila The morphology of jaws and teeth, and to a and Richardsonius); molariform as well as lesser extent, body size and shape, are often grinding indicate omnivory used with data from recent analogues to (Idadon); nocturnal scavenging and omnivory indicate predator-prey relationships and feed- is inferred by analogy to similar modern ing behaviors. For example, fish in the Green relatives (Ictalurus); a medium-sized predator is River fauna have been ecologically classified inferred by analogy to a similar modern 581 relative (Archoplites); small benthic predators principles except that transport may occur on crustaceans are inferred by analogy to when carcasses float (Sch/ifer, 1972). Even this similar modern relatives (Cottidae). Some tro- principle has been ignored or used inappropri- phic inferences are complex: in Oncorhynchus, ately because of misconceptions regarding its numerous gill rakers indicate planktivory, but physical basis and conditions of occurrence. in the same form, piscivory is inferred from Students of fish paleolimnology have relied analogy with relatives; in Prosopium, rela- upon traditional but questionable explana- tively large jaws and large size indicate tions for '~good preservation": lack of scaven- piscivorous habits, but also planktivory is gers, rapid , anaerobic condi- inferred by analogy with relatives. Smith, G.R. tions, and, since Sch/ifer's work, occasional and Todd (1984) extended these inferences to flotation. Nevertheless, aquatic taphonomy generalizations about broader aspects of the has been in the forefront of experimental trophic interactions and the resulting selective discovery of processes of decay and burial, pressures for defensive spines in Lake Idaho primarily through the work of Sch/ifer and sculpins, other German workers on ~Aktuopal~iontolo- The use of analogy is more acceptable when gie" (e.g. Richter, 1928; Sch/ifer, 1964, 1976). similar congeners are being compared, such as Aquatic taphonomy had its beginnings in the within Amia and within Lepisosteus in the work of Weigelt (1927) and Hecht (1933) who and Recent, and in most fish genera discussed processes contributing to decay and from the Pliocene, Pleistocene, and Recent. In fossilization of many types of organisms, in- most cases, however, the authors have been too cluding fish. The classic work on mass mor- willing to infer specialized feeding habits from tality of fishes and the implications of this specialized morphology. Ecological studies on phenomenon for the fossil record is that of modern specialists indicate extensive omni- Brongersma-Sanders (1948, 1949, 1957). Gun- vory, especially in '~" and molluski- ter's observations (1941, 1942, 1947) of mass vores. Because digestion of plant material is death on the Gulf Coast of North America also not enzymatically efficient in fishes, and pro- contribute to this area of study. Both Bron- tein is too valuable to exclude from the diet, gersma-Sanders and Gunter dealt almost exclu- opportunistic feeding is probably selected for sively with mass death in . in fishes that can eat plants. Experimental Gunter pointed out (1948) that many examples studies (Liem, 1980) indicate that even morpho- of mass death (inferred from multiple indi- logical specialists may be behavioral opportun- viduals on single bedding planes) are found in ists. with high organic content and in areas of probable former upwelling. This corre- The use of taphonomy sponds to Brongersma-Sanders' observations of recent mass death episodes caused by toxicity Although taphonomy as a field of study has of blooms in offshore areas of enjoyed increasing popularity in the past 10 upwelling. Gunter's observations were of near- , recent summaries (Dodson, 1980; Olson, shore shallow habitats, in which fish and other 1980) of the current state of the discipline do marine organisms were killed by sudden tem- not discuss the taphonomy of the inhabitants perature changes, for example, the swift pas- of aquatic systems. Most workers have concen- sage of a cold front. He noted (1947) that these trated on fluvial transport and burial of episodes of mass death may be recorded in the from terrestrial , or the fossil record if nearby runoff is sufficiently transport and sorting of marine invertebrates, strong to cause sedimentation or if low temper- Unlike studies of transport, sorting, and abra- atures or strong retard decay. sion based on hydrodynamic flow, studies of The most important general work in aquatic quiet water systems have had few guiding taphonomy has been that of Sch~ifer (1972) who 582

stressed the importance of directly observing soupy ooze over a hard surface presumably processes of death, decay, and burial. His formed by synsedimentary compaction. In this monograph, Ecology and Paleoecology of model, the carcass would settle down in the Marine Environments, serves as a seminal soft ooze to the hard surface, where very work in promoting experimental approaches to quickly the sediments would compact around it paleoecology. Investigating the preservation of before decay could occur. We would interpret North Sea fishes, he noted flotation in re- the specimens to have dropped into deep water sponse to internal gas production and com- after floating and decaying at the surface. mented on the loss of body parts from floating Partial articulation is attributable to unde- carcasses, cayed connective tissue. The observed spaces Other works that describe the preservation were probably filled with tissue at the time of of fish in ancient waters include Dence (1956), burial. The rationale for this alternate inter- Breder (1957), Raynor (1963), Wilson (1968), pretation is developed in the sections to follow. Cavender (1969), Melton (1970), Perkins (1970b), Richardson and Johnson (1971), Zan- Eocene gerl (1971), Buchheim and Surdam (1977), In astudy of lacustrine sediments, Buchheim Wilson (1978), Boyer (1981, 1982), and Franzen and Surdam (1981) drew on several lines of et al. (1982). David's (1957) annotated bibli- evidence including sedimentary structures, ography is a rich source of earlier works on plant remains, insects, aquatic invertebrates, fish preservation, paleoecology and mass mor- and paleoecology of the Green River Eocene tality, fish fauna to infer two major paleoenviron- ments (nearshore and offshore) of Lake Gosi- Environment of ute. The lake was interpreted as shallow with oscillating water levels, both on geological marine shales grounds (alternating sedimentary packages) In the first comprehensive case history of and on paleoecological grounds (alternating fish preservation, Zangerl and Richardson times of littoral and limnetic communities). (1963) studied the fish fauna of carbonaceous The nearshore littoral habitat is represented Pennsylvanian shales from the Illinois Basin of by shales displaying oolites, horsetail and the United States. The authors were led by cattail fragments, terrestrial insects, ostra- common assumptions about preservation to , , and fish fry. These shales unusual conclusions about the environment, are laminated, sometimes papery, and siliceous From restricted assumptions about causes of with aragonite or calcite. No or good preservation and disarticulated car- root mottling was observed. The fry and casses, they concluded that the shales were juvenile fish are almost all Knightia and are deposited in a lagoon covered by a floating mat very well preserved. The authors suggest that of vegetation. To explain the disarticulated the finely preserved fish of this shallow-water skeletons, they hypothesized feeding frenzies area "...probably were seeking the protection beneath this mat and common regurgitation of of the shallows and aquatic vegetation..." prey items. Under these circumstances, the (Buchheim and Surdam, 1981, p. 425) and that excellent preservation required sedimentation the cause of death was alkaline/saline shock. rates to be several orders of magnitude greater Figured examples (Buchheim and Surdam, than any known from similar depositional 1981, fig.14-6E, F) show that the fish had jaw systems today (Sadler, 1981). But, to explain tetany, which is also known to occur from heat why carcasses were not totally disarticulated shock or anoxia. and why internal spaces of the bones were not There are several problems with the in- filled with during decay, the bottom terpretation of this assemblage as indicative was inferred to have been a combination of of a shallow water habitat. First, littoral areas 583 exhibit the greatest diversity within lakes, yet turnover) could scavengers venture to the in this example, littoral diversity is low bottom. The alternation of occasionally scav- compared to the limnetic diversity. Second, enged mass-death layers and barren intervals the sediments are unbioturbated even though supports this hypothesis. the waters were interpreted to be warm and If the stratified lake model were to be shallow- and therefore hospitable to bottom- rejected on other grounds, it would still be disturbing vegetation, benthic invertebrates, necessary to explain why the fish killed by and fish. The excellent preservation of the saline shock in a shallow playa lake system did fish is not consistent with warm shallow not float in the warm dense waters. This choice waters inhabited by scavengers. This point between lake models is discussed further in the will be discussed in more detail in the next sections on the Stratified Lake Model and section. Overturn and Fish Mortality. The offshore limnetic (but not deepwater) habitat was characterized by the presence of Temperature,bacterial decay, and flotation well-laminated carbonate rocks containing a typical Green River fish fauna. Mass death Experiments show (Smith and Elder, 1985), layers of Knightia are common, especially in following Sch/ifer (1972), that the partial disar- the lower 10 m of the measured sections. These ticulation of skeletons such as observed by mass are explained as the result of Zangerl and Richardson (1963) is the result of change in alkalinity/salinity conditions with bacterial decay and flotation in warm waters. zones barren of fish interpreted as indicating Such carcasses lose elements and body regions extreme conditions. The authors suggest that piecemeal before sinking to the bottom as the Amyzon was a along with the gas escapes. The observed dispersion of ele- catfish Astephus and that both were scaven- ments for short distances in several directions gers on the mass death carcasses. However, from bottom carcasses (explained as predation this interpretation is not consistent with the injury) may indicate some scavenging activity functional morphology of Amyzon, which had (see below). Abnormal environmental condi- a terminal, not ventral, mouth and was prob- tions are not required to preserve these car- ably a midwater feeder on . Ac- casses in the observed condition. Sufficient cording to Buchheim and Surdam the numer- conditions are warm, relatively shallow water, ous observed are attributable to the occasional scavengers, and reasonable rates of catfish Astephus and show that it was a bottom sedimentation. resident of this habitat when lake conditions Smith, G. R. and Elder (1985) have shown were sufficiently fresh and oxygenated, that the undisturbed fossil fish of the Clarkia But the limited evidence for scavenging, of Basin in Idaho were not buried for at least either the mass death layers of Knightia or of several months after death. Other instances of Astephus, indicates limited oxygen conditions this phenomenon can be observed in both the (see below) or bottom waters too saline or fossil record and recent lake bottoms. For otherwise toxic to support popula- example, skeletons resting uncovered at sev- tions. Coprolites, not necessarily of Astephus, eral meters depth and less than 15°C on the need only be evidence of surface water popula- bottom of Lake Michigan have been observed tions, to remain intact for several weeks though the This model of a differentiated water column flesh became partially decayed (J. A. Dorr, III, is the same as Bradley's (1948) stratified lake pers. comm., 1984). Carcasses may be moved by with a saline and/or anaerobic hypolimnion, currents and scavenged by crayfish. and is consistent with the ideas of Desborough A pertinent contribution to this aspect of (1978) and Boyer (1981, 1982). In this model, aquatic taphonomy is Waldman's (1971)study only rarely (such as during times of lake of the Koonwarra freshwater fauna from the 584

Lower of southeast Australia. Fol- The stratified lake model lowing the traditional assumptions about good There is, however, much evidence at Koon- preservation, he reached conclusions that bear warra in favor of a stratified lake. In addition reconsideration in light of experiments on to the presence of undisturbed, varved sedi- temperature and flotation. Waldman envi- ments, Waldman also noted that there is little sioned a shallow arm of a lake that was cut off evidence for bacterial decay and gas rupture of every winter by a sheet of cloudy ice, under the thoracic cavities. Experimental results of which the waters became anoxic; the fish Smith, G. R. and Elder (1985; Elder, 1985) (including ) died and their carcasses indicate that temperature and pressure control were covered by clay settling out during the the buoyancy of decaying fish through regula- winter months, tion of bacterial gas production, and that sup- Because of the lack of scattering by scaven- pression of gas bubbles is possible only in cold gers, Waldman saw a need to postulate a or deep waters, where bacterial is mechanism for covering the carcasses and limited and bubbles are small. The observation thus preventing disturbance by other - of completely articulated fossil fish suggests isms. Ice cover was the only hypothesis that cold or deep waters in which the fish were seemed to explain the lack of disturbance: prevented from floating. Other hypotheses -- floating vegetation mat, These results suggest that the Koonwarra salt crust, organic scum, or oil film -- were fish carcasses, unlike the fishes in Zangerl and considered and rejected because no evidence Richardson's Pennsylvanian shales, were not of their presence was found in the record, warm enough or shallow enough to float. Waldman noted varved sediments that would Waldman noted that isolated elements and normally indicate a seasonal cycle of stratifi- partial carcasses are rare. If cold or deep cation of the water body. But he rejected the waters are accepted on the basis of similarity possibility that an oxygen deficit at the bot- to experimental preservation in cold water, tom of a deep stratified lake protected the then the most plausible model for death and carcasses from scavengers (see below), be- burial is that of a stratified lake in which fish cause of a previous conclusion that the fish settled into cold or deep waters, where they and associated fauna dwelt only in were rarely disturbed by scavengers because of shallow waters. He assumed that they could oxygen depletion. not have inhabited (or withstood transport Wilson (1977, 1980)studied lacustrine sedi- to) the center of a deep lake. Therefore, ments and fishes in Eocene deposits of British despite paleoclimatic evidence and the Columbia, Canada. Integrating climatic evi- analogy of warm-water habitats of today's dence with that from fishes, insects, varves, lungfish (Schaeffer, 1970), Waldman hypothe- and flora he inferred death from circulation sized cold waters and cold winters. ('~turnover") of a stratified, warm monomictic Reviewing the evidence, Wilson (1977) noted lake. Good preservation was attributed to several inconsistencies in Waldman's hypothe- exclusion of scavengers from an anoxic hypo- ses. Wilson suggested that in shallow waters, limnion. warmer temperatures of the following spring should have caused decay and flotation of Overturn and fish mortality the winter-killed fish. Furthermore, although McGrew's (1975) perceptive analysis of death flocculated, ~'glue-like", clay could have and burial of fishes in the Eocene Green River trapped fish, varves such as those seen in Formation, citing Bradley's earlier work Waldman's section are only formed by un- (1948),hypothesized that the fish died primar- flocculated materials. These arguments sug- ily from toxicity generated by summer algal gest that the shallow ice cover model is blooms in a stratified lake. The stratified lake untenable, model has recently been challenged by Surdam 585 and various coworkers (Eugster and Surdam, But unburied fish will not remain on the 1973; Lundell and Surdam, 1975; Surdam and bottom unless cold or under pressure, and the Wolfbauer, 1975; Buchheim and Surdam, 1977; amount of disarticulation depends primarily Surdam and Stanley, 1979) but has received on other factors. In the Fossil Lake fishes, the support from Desborough (1978) and Boyer disarticulation patterns are mostly consistent (1981, 1982). The Green River lake system may with the scavenging preferences of the larger have included both types of lakes; it comprised such as snails and crayfish that several separate basins of varying morphome- prefer the head and gut cavities to the muscu- try. Thus Lake Uinta and Lake Gosiute may lar areas of the body (see below). have been playa lakes or lakes showing Perkins (1970b) favored a model in which ectogenic meromixis (Boyer, 1981, 1982) or unseasonably low temperatures in summer biogenic-chemical stratification (Desborough, cause circulation of toxic H2S from the hypo- 1978). The following discussion is based pri- limnion of a normally dimictic lake. An marily on the preservational characteristics of alternate interpretation may apply to Fossil fishes in Fossil Lake, which supported fish Lake -- a warm monomictic lake with winter diversity that we do not believe to be possible circulation in a warm-temperate to subtropical in a playa lake. climate. The warm monomictic model assumes McGrew was the first to attempt to determine moderate . The fish would be season of death based on preservation pattern, killed whenever circulation brought hydrogen [In Waldman's (1971) hypothesis, season of sulphide into the epilimnion from the hypolim- death was a corollary of his hypothetical cause nion as the lake cooled in winter. The fish of death.] McGrew noted that it is common to would settle to the bottom in cold water and be find the greatest disarticulation in the head and disturbed by scavengers when sufficient oxy- anterior half of the body in Fossil Lake fishes, gen was present. In spring the surface waters The sequence of disarticulation was estimated would warm up and stratify once more. The to proceed from the anteromost vertebrae for- hypolimnion would remain cold and spring ward to the head and backward to the tail at algal growth would cause carbonate precipita- different rates. The most posterior part of the tion accompanied by oxygen depletion. In deep body is commonly undisturbed. This was ex- waters, carcasses could be covered completely plained as the result of easy access for bacteria to lie without further disturbance; in shallower in the gill region and orifices of the head. areas, they would be exposed to further scav- Noting that the fishes in the mass mortality enging or flotation. The biased nature of the layers are not disarticulated, McGrew pro- FossilLake collections, towardsthe occasional posed that mortality and preservation were deep-lake strata with well-articulated indi- caused by one event and suggested that rapid viduals, should be recognized. Flotation and burial following an algal bloom would be one scattering of isolated elements may have been way to accomplish this. [In lakes, circulation more frequent than represented in museums. of water and nutrients may be followed by an algal bloom that depletes the water of carbon Seasonality dioxide, causing precipitation of carbonates. Three tests of season of death may be done, The algae may also release metabolites that disarticulation sequence, seasonal bone may become toxic to fishes (Prescott, 1948).] growth (Wilson, 1977; Smith, 1985; Elder, 1985), Postulating a summer algal bloom, fish mor- and population structure (Wilson, 1977). One tality, and a sequence of decay and disarticula- can extend McGrew's principles, sequencing tion of carcasses, he suggested that the amount the disarticulation and then calibrating the of disarticulation is directly proportional to rate of disarticulation using experimental re- the amount of time the fish lay on the bottom sults, obtained under different conditions of before the next carbonate precipitation event, temperature, current energy, and scavengers. 586

Experiments show that at low temperatures interpreted with the help of some principles (4-6°C) with no current energy or scavengers, from information theory, aids in the discrimi- lepidotrichia (fin ray bones) require at least nation. Viewed in the framework of informa- four weeks to disarticulate (Smith, G. R. and tion theory, taphonomy attempts to retrieve Elder, 1985). One may estimate a minimum environmental information from the pattern of time from death to burial and check this loss of organic organization after death and against the hypothetical sequences of lacus- the new information added to the system by the trine events, taphonomic processes themselves (Tasch, Alternatively, one can examine individual 1965). As the processes of death, decay, and bones and scales for growth annuli (McGrew, burial destroy information about the , 1975; Wilson, 1977; Smith, G. R., 1985). Season they add information about its organic and of death will be recorded by the place in the physical environment and its history in that cycle where growth stopped. Thus, a fish bone environment. whose edge displays a dense band with no rapid A first step towards quantifying the informa- growth beyond it indicates an individual that tion in a dispersed skeleton is the measurement overwintered (at least partially) but which did of distance between bones with known initial not survive through the following spring. An relationship, e.g. paired bones. The correlation incomplete light band of rapid summer growth between positions of members of pairs and the indicates summer mortality, frequency distribution of the distances and di- Wilson (1977) used standard length as a rectionality of transport of series of pairs allow proxy for body size and noted that a histogram us to use information theory to measure attri- of the most numerous , the sucker butes of transport and burial processes, such as Amyzon, showed distinct polymodality in an direction and heterogeneity of the energy re- Eocene lake in British Columbia. This sug- sponsible for transport and the rate of burial. gested that death was catastrophic and occur- Experiments with scavengers (e.g. gastro- red at a single time each , thus sampling pods; Elder, 1985) indicate that fish bones are distinct year-classes. This yearly event fits his normally moved in all directions and are with- model postulating a warm monomictic lake out preferred orientation (Fig.2). The distance with winter circulation. Analysis of Wilson's of displacement is not correlated between preservational patterns (Elder, 1985) in light paired bones nor does it seem correlated with of the experimental demonstration of fish flota- the hydrodynamic character of the bone. Ob- tion only at higher temperature and low pres- servation of this pattern in the fossil record sure (Smith, G. R. and Elder, 1985) and patterns indicates scavenger effects (Fig.3). Its absence caused by scavenging and currents (Elder and may indicate conditions unsuitable for scaven- Smith, G. R., 1984) confirm Wilson's conclu- gers, such as anoxic bottom waters. Experi- sions, ments show that anoxic waters do not prevent decay or flotation, contrary to Sch~fer (1972), Disarticulation by scavengers and transport by but can preclude scavenging. currents. Experiments with currents (Elder, 1985) show a more unimodal distribution of dispersal The importance of evidence of disturbance directions. Often the elements exhibit a pre- has been emphasized in the previous discussion ferred orientation. There is a correlation be- of paleoenvironmental studies. Because pres- tween distances moved by members of pairs ence or absence of scavengers may provide and between elements having similar shapes. A information about oxygen and other important set of bones used in current experiments is aspects of lakes, it is necessary to distinguish diagrammed in Fig.4 with histograms of log among possible causes of skeletal disturbance difference in distance moved. The distribution and rearrangement. Experimental evidence, shows, for example, that in 22 out of 30 587

=ku,, ~

...... :

pl c~ I

Fig.2. Distribution of major bones of a sunfish, ca. 5 cm long, after 14 days of scavenging by eight 5 12 mm gastropods. The fish's starting position is outlined on the left. Identification of bones drawn: A=angular; C=; D=dentary; E= ectopterygoid; H= ; I = interopercle; L = lacrimal; M= maxilla; N= pharyngeal; O = opercle; P = preoper- cle; Pro= premaxilla; Q= quadrate; R= ceratohyal; S=subopercle; Skull= frontals and parietals; T= supratemporal; U = urohyal; V = vomer.

Fig.3. Priscacara sp., Green River Formation, Fossil, Wyoming, University of Michigan Museum of Paleontology V56623, ca. 14 cm. Note multidirectional scatter of bones and scales, especially from the head and abdomen, indicating scavenger transport. 588

experimental trials the pharyngeal bones s iLiL travelled together. This distribution is gener- i~ . ~~L .~ i~ s~L ally the same for most of the elongate elements ~ ~ and those of more 3-dimensional shape, such as ~" ~ the frontals. Some of the highly 3-dimensional bones do not show this distribution so clearly because they are hydrodynamically more sta- ble only in certain positions. Members of the IL ~L L i[ [ L pair become set in the stable position at differ- ~ ent times and thus become more separated s than simpler bones. The flat bones of the oper- ...... ~ ~ ...... ! ..... | ~ ...... cular series show a random pattern. Fossil fish may be interpreted as having been under a flowing current when we observe, for example, directional scatter of scales or lepido- trichia (Fig.5), or when we observe clear indica- il IL L L !~ ![ tionofcorrelationofdistancesmovedbydisarti- ~= ~ ~ culated, paired elongate elements. Evidence for ..... | ..... | ~ ...... ~s ...... | ..... : gentle, directional, and relatively homogeneous 0 5 disturbance (in unburrowed sediments) also im- LOGDISTANCE BETWEEN MEMBERS OF PAIREDBONES AFTER TRANSPORT plies that the fish was not buried during the time Fig.4. Frequency distributions of distances (log cm) of decay (Smith, G. R. and Elder, 1985). From between paired bones after transport by currents in an measurements of decay of fish under different experimental tank. Certain shapes, especially elongate conditions, we may infer how long a fossil fish bones, show correlated responses to current, providing remained uncovered, an aspect of sedimentation taphonomic information. Thirty trials represented, rate.

Fig.5. Plioplarchus sp., Sentinel Butte, North Dakota, University of Michigan Museum of Paleontology V83213, ca. 10 cm. Scales and lepidotrichia show nearly unimodal dispersal and correlated positions of similar elements, indicating current transport. 589

Evidence for the action of both processes is cesses in which information is lost, reorgan- found in those specimens exhibiting multi- ized, or stored under the control of biological directional scatter and loss of many hydrody- activity or hydrodynamic energy. Because namically light elements, such as lepidotrichia potential taphonomic information is correlated and scales, from the area. A scavenger that with the structural complexity of the organ- disturbs a skeleton after current activity may isms, fish skeletons are rich sources of data. destroy the characteristic pattern of correla- Information loss is proportional to the number tion and cause an overprint of the multi- of ways in which parts of the system (various directional scatter pattern. In contrast, a bones) can be taphonomically arranged, and current acting after scavenging would not be increases as physical factors such as time, likely to create a pattern of distance correla- energy, chemical reactions, and scavengers tion among already scattered bones. This may increase the system's entropy. The several allow distinction of the sequence of distur- distinctive ways in which parts of the fish bance, skeleton may be rearranged (e.g. by bacteria, currents, or scavengers) provide new informa- Discussion tion.

Interpretation of fossil skeletal patterns of Acknowledgements disarticulation, disorientation, and displace- ment (according to the results of observable The authors would like to thank Bill Pelle- aquatic processes) allows the possibility of tier for assistance with photography. determining temperature and depth of lakes, inferring types of lake stratification (if any) References and seasons and causes of death, and of estimating oxygen conditions and something of Barlow, G. W., 1961. Causes and significance of morpholo- the and rate of sedimentation. A gical variation in fishes. Syst. Zool., 10:105 117. Bell, M. A., 1973. Pleistocene three-spine sticklebacks, major advantage of using a taphonomic ap- Gasterosteus aculeatus, (Pisces) from southern Califor- proach to paleoecology is its testability. There nia. J. Paleontol., 47(3): 479-483. are two approaches. The traditional one has Bennett, D. K., 1979. Three Late Cenozoic fish faunas from been the comparative approach to preservation Nebraska. Trans. Kans. Acad. Sci., 82(3): 146 177. Boyer, B. W., 1981. Tertiary lacustrine sediments from systems, regarding them as natural experi- Sentinel Butte, North Dakota, and the sedimentary ments in which some factors were kept con- record of ectogenic meromixis. J. Sediment. Petrol., stant and others varied. This approach may be 51(20): 429-440. misleading when examples are chosen to illus- Boyer, B. W., 1982. Green River laminites: Does the playa- lake model really invalidate the stratified-lake model? trate conclusions drawn previously. Geology, 10(6): 321 324. On the other hand, the experimental ap- Bradley, W. H., 1948. Limnology and the Eocene lakes of proach attempts to test taphonomic hypotheses the Rocky Mountain region. Geol. Soc. Am. Bull., 59: using Recent analogies in controlled systems 635-648. Breder, Jr., C.M., 1957. A note on preliminary stages in the (Schiifer, 1972). Ancient systems may then be fossilizationof fishes. Copeia, 1957(2): 132 135. interpreted by the similarities of their preser- Brongersma-Sanders, M., 1948. The importance of upwell- vation to results obtained by known processes, ing to and oil geology. Verh. K. involving factors such as time, energy, and Ned. Akad. Wet., Afd. Nat. Sect. 2, 45(4): 1 116. Brongersma-Sanders, M., 1949. On the occurrence of fish water, remains in fossil and recent marine deposits. Bijdr. Experimental studies offer the surest way to Dierk., 28: 65-76. discover reliable generalizations about preser- Brongersma-Sanders, M., 1957. Mass mortality in the sea. vation of aquatic as well as terrestrial organ- In: J. W. Hedgpeth (Editor), Treatise on Marine Ecology and Paleoecology, I: Ecology. Geol. Soc. Am. Mem., 67: isms in lakes and rivers. Decay, transport, and 941-1010. burial can be experimentally studied as pro- Buchheim, H. P. and Surdam, R. C., 1977. Fossil catfish and 590

the depositional environment of the Green River Forma- Gaudant, J., 1979a. Principes et m~thodes d'une pal~o- tion, Wyoming. Geology, 5(4): 196-198. ichthyologie bathym~trique. Palaeogeogr., Palaeoclima- Buchheim, H. P. and Surdam. R. C., 1981. Paleoenviron- tol., Palaeoecol., 28:263 278. ments and fossil fishes of the Laney Member, Green Gaudant, J., 1979b. Pal~oichthyologie et pal~osalinit~s River Formation, Wyoming. In: J. Gray, A. J. Boucot and propos de certains d~pots Cenozoiques de la M~diterra- W. B. N. Berry (Editors), Communities of the Past. n~e occidentale. In: 7e R~un. Annu. Sci. Terre, 1979, Dowden, Hutchinson and Ross, Stroudsburg, Pa., p. 208. pp. 415-452. Gaudant, J., 1981. L'ichthyofaune de Messinien continen- Cavender, T. M., 1969. An mudminnow (Family tal d'Italie septentrionale et sa signification g~odynami- Umbridae) from Oregon with remarks on relationships que. Palaeontographica A, 172: 72-102. within the Esocoidei. Occas. Pap. Mus. Zool., Univ. Grande, L., 1980. Paleontology of the Green River Forma- Mich., 660: 1-33. tion, with a review of the fish fauna. Geol. Surv. Wyo. Chaney, R. W., Condit, C. and Axelrod, D. I., 1944. Pliocene Bull., 63: 1-333. floras of California and Oregon. Carnegie Inst. Wash. Gunter, G., 1941. Death of fishes due to cold on the Texas Publ., 553: 1-407. coast, January 1940. Ecology, 22: 203-208. Cross, F. B., 1970. Fishes as indicators of Pleistocene and Gunter, G., 1942. Offats Bayou, a locality with recurrent Recent environments in the Central Great Plains. In: summer mortality of marine organisms. Am. Midl. Nat., W. Dort, Jr. and J. K. Jones (Editors), Pleistocene and 28: 631-633. Recent Environments of the Central Great Plains. Univ. Gunter, G., 1947. Catastrophism in the sea and its Kans. Spec. Publ., 3: 241-258. paleontological significance, with special reference to David, L. R., 1944. Reliability in determination of bathy- the Gulf of Mexico. Am. J. Sci., 245(11): 669-676. metric conditions of deposition based on fossil fish Hecht, F., 1933. Der Verbleib der organischen Substanz der Tiere bei meerischer Einbettung. Senckenbergiana, remains. Geol. Soc. Am. Bull., 55: 1467-1468. David, L. R., 1957. Fishes (other than ). In: H.S. 15(3 4): 165-249. Hibbard, C. W., 1955. The Jinglebob interglacial (Sango- Ladd (Editor), Treatise on Marine Ecology and Paleo- mon?) fauna from Kansas and its climatic significance. ecology, II: Paleoecology. Geol. Soc. Am. Mem., 67: Contrib. Mus. Paleontol., Univ. Mich., 12: 179-228. 999-1010. Hibbard, C. W., 1960. An interpretation of Pliocene and Dence, W. A., 1956. Concretions of the alewife, Pomolobus Pleistocene climates in North America. Annu. Rep. (Wilson), at Onondaga Lake, New York. pseudoharengus Mich. Acad. Sci. Arts Lett., 62: 5-30. Copeia, 1956(3): 155 158. Hibbard, C. W. and Dalquest, W. W., 1966. Fossils from the Desborough, G. A., 1978. A biogenic-chemical stratified Seymour Formation of Knox and Baylor Counties, lake model for the origin of oil shale of the Green River Texas, and their bearing on the Late Kansan climate of Formation: an alternative to the playa-lake model. Geol. that region. Contrib. Mus. Paleontol., Univ. Mich., 21(1): Soc. Am. Bull., 89: 961-971. 1-66. Dodson, P., 1980. Vertebrate burials. , 6(1): Johnson, R. K. and Barnett, M. A., 1975. An inverse 6 8. correlation between meristic characters and food supply Dorf, E., 1930. Pliocene floras of California. Carnegie Inst. in mid-water fishes: evidence and possible explanations. Wash. Publ., 412: 1-108. Fish. Bull., 73: 284-298. Dorf, E., 1970. Paleobotanical evidence of Mesozoic and Liem, K. F., 1980. Adaptive significance of intra- and Cenozoic climatic changes. In: E. L. Yochelson (Editor), interspecific differences in the feeding repertoires of Proc. North American Paleontological Convention, Part cichlid fishes. Am. Zool., 20: 295-314. D, . Allen Press, Lawrence, Kans., pp. Lindsey, C. C. and Arnason, A. N., 1981. A model for 323-346. responses of vertebral numbers in fish to environmental Elder, R. L., 1985. Principles of Aquatic Taphonomy with influences in development. Can. J. Fish. Aquat. Sci., 38: Examples from the Fossil Record. Thesis. Univ. Mich- 334 347. igan, Ann Arbor, Mich., 336 pp. Lundberg, J. G., 1967. Pleistocene fishes of the Good Creek Elder, R. L. and Smith, G. R., 1984. Fish taphonomy and Formation, Texas. Copeia, 1967(2): 453 455. paleoecology. G~obios M~m. Sp~c., 8:287 291. Lundell, L. L. and Surdam, R. C., 1975. Playa lake Eshelman, R. E., 1975. Geology and paleontology of the deposition, Green River Formation, Piceance Creek Early Pleistocene (Late Blancan) White Rock fauna Basin, Colorado. Geology, 3: 493-497. from north-central Kansas. Pap. Paleontol., 13: 1-60. McGrew, P. O., 1975. Taphonomy of Eocene fish from Eugster, H. P. and Surdam, R. C., 1973. Depositional Fossil Basin, Wyoming. Fieldiana Geol., 33(14): 257-270. environment of the Green River Formation of Wyo- Melton, W. G., Jr., 1971. The Bear Gulch fauna from central ming, a preliminary report. Geol. Soc. Am. Bull., 84: Montana. In: E. L. Yochelson (Editor), Proc. North 1115 1120. American Paleontological Convention, Part I, Extra- Franzen, J. L., Weber, J. and Wuttke, M., 1982. Sencken- ordinary fossils. Allen Press, Lawrence, Karts., pp. berg-Grabungen in der Grube Messel bei Darmstadt, 3: 1202 1207. Ergebnissse 1979-1981. Cour. Forschungsinst. Sencken- Miller, R. R., 1965. freshwater fishes of North berg, 54:1 118. America. In: H. E. Wright, Jr. and D. G. Frye (Editors), 591

The Quaternary of the United States. Princeton Univ. Semken, H. A., Jr., 1966. and paleontology of Press, Princeton, N.J., pp. 569-581. the McPherson Equus beds (Sandahl local fauna), Myers, G. S., 1949. Salt-tolerance of fresh-water fish groups McPherson County, Kansas. Contrib. Mus. Paleontol., in relation to zoogeographical problems. Bijdr. Dierk., Univ. Mich., 20(6): 121-178. 28:315 322. Sherrod, N., 1963. Late Pleistocene fish from lake sedi- Neff, N. A., 1975. Fishes of the Kanopolis local fauna ments in Sheridan County, North Dakota. N. D. Acad. (Pleistocene) of Ellsworth County, Kansas. In: G.R. Sci. Proc., 17: 32-36. Smith and N. E. Friedland (Editors), Studies on Cenozoic Smith, C. L., 1954. Pleistocene fishes of the Berends fauna Paleontology and Stratigraphy. Pap. Paleontol., 12: of Beaver County, Oklahoma. Copeia, 1954(4): 282-289. 39-48. Smith, C. L., 1958. Additional Pleistocene fishes from Olson, E. C., 1980. Taphonomy: its history and role in Kansas and Oklahoma. Copeia, 1958(3): 176-180. community evolution. In: A. K. Behrensmeyer and A.P. Smith, C. L., 1962. Some Pliocene fishes from Kansas, Hill (Editors), Fossils in the Making: Vertebrate Oklahoma and Nebraska. Copeia, 1962(3): 505 520. Taphonomy and Paleoecology. Univ. Chicago Press, Smith, C. L., 1964. Fishes and climates. Nat. Hist., 73(2): Chicago, Ill,, pp. 5 19. 34-39. Ossian, C. R., 1973. Fishes of a Pleistocene lake in South Smith, G. R., 1963. A Late Illinoian fish fauna from Dakota. Publ. Mus. Mich. State Univ., Paleontol. Ser., southwestern Kansas and its climatic significance. 1(3): 101 124. Copeia, 1963(2): 278 285. Perkins, P. L., 1970a. Equibility and trophic levels in an Smith, G. R., 1975. Fishes of the Pliocene Glenns Ferry Eocene fish population. Lethaia, 3(3): 301 310. Formation, southwest Idaho. Pap. Paleontol., 14: 1-68. Perkins, P. L., 1970b. Fossil Lake, Wyoming, and its Smith, G. R., 1978. of intermountain fishes. magnificent inhabitants. Discovery, 6(1): 33 40. In: K. T. Harper and J. L. Reveal (Editors), Intermoun- Prescott, G. W., 1948. Objectionable algae with reference tain Biogeography: A Symposium. Great Basin Mem., 2: to the killing of fish and other animals. Hydrobiologia, 1: 17 42. 1 13. Smith, G. R., 1981. Effects of habitat size on species Radforth, I., 1944. Some considerations on the distribution richness and adult body sizes of desert fishes. In: R. J. of fishes in Ontario. Contrib. R. Ont. Mus. Zool., 25: Naiman and D. L. Soltz (Editors), Fishes in North 1 116. Raynor, D. H., 1963. The Achanarras Limestone of the American Deserts. Wiley, New York, N.Y., pp. 125-171. Middle Old Red Sandstone, Caithness, Scotland. Proc. Smith, G. R., 1985. Paleontology of Hidden Cave: fish. In: Yorks. Geol. Soc., 34(2): 117-138. D.H. Thomas (Editor), The Archeology of Hidden Cave. Richardson, E. S., Jr. and Johnson, R. G., 1971. The Mazon Anthropol. Pap. Am. Mus. Nat. Hist., 61:171 178. Creek fossils. In: E. L. Yochelson (Editor), Proc. North Smith, G. R. and Elder, R. L., 1985. Environmental American Paleontological Convention, Part I, Extra- interpretation of burial and preservation of Clarkia ordinary fossils. Allen Press, Lawrence, Kans., pp. fishes. In: C. J. Smiley (Editor), Late Cenozoic History 1222 1235. of the Pacific Northwest. Pac. Div. Am. Assoc. Adv. Sci., Richter, R., 1928. Aktuopal/iontologie und Pal~iobiologie, Calif. Acad. Sci., San Francisco, Calif., pp. 85 93. eine Abgrenzung. Senckenbergiana, 10:285 292. Smith, G. R. and Lundberg, J. G., 1972. The Sand Draw fish Rogers, K. L., Repenning, C. A. et al., 1985. Middle Pleisto- fauna. In: M. F. Skinner and C. W. Hibbard (Editors), cene (Late Irvingtonian: Nebraskan) climatic changesin Early Pleistocene pre-glacial and glacial rocks and south-central Colorado. Nat. Geogr. Res., 1:535 563. faunas of north central Nebraska. Bull. Am. Mus. Nat. Sadler, P. M., 1981. Sediment accumulation rates and the Hist., 148(1): 40 54. completeness of stratigraphic sections. J. Geol., 89: Smith, G. R. and Todd, T. N., 1984. Evolution of species 569-584. flocks of north-temperate lacustrine fishes. In: A. A. Schaeffer, B., 1970. Mesozoic fishes and climate. In: E.L. Echelle and I. Kornfield (Editors), Yochelson (Editor), Proc. North American Paleontologi- Species Flocks. Univ. Maine at Orono Press, Orono, cal Convention, Part D, Paleoclimatology. Allen Press, Maine, pp. 45-68. Lawrence, Kans., pp. 376-388. Smith, M. L., 1981. Late Cenozoic fishes in the warm Sch/ifer, W., 1972. Ecology and Palaeoecology of Marine deserts of North America: a reinterpretation of desert Environments. Oliver and Boyd, Edinburgh, 568 pp. . In: R. J. Naiman and D. L. Soltz (Editors), (translated by I. Oertel). Fishes in North American Deserts. Wiley, New York, Sch~ifer, W., 1964. Aktuopal~iontologische Beobachtungen. N.Y., pp. 11 38. 2. Zur Fossilisation von Limulus. Nat. Mus., 94(3): Smith, M. L., Cavender, T. M. and Miller, R. R., 1975. 98-104. Climatic and biogeographic significance of a fish fauna Sch~ifer, W., 1976. Aktuopal/iontologische Beobachtungen. from the Late Pliocene-Early Pleistocene of the Lake 10. Zur Fossilisation von Vogeln. Nat. Mus., 106(9): Chapala Basin (Jalisco, Mexico). In: G. R. Smith and 276-- 279. N.E. Friedland (Editors), Studies on Cenozoic Paleonto- Schultz, G. E., 1965. Pleistocene vertebrates from the logy and Stratigraphy. Pap. Paleontol., 12:29 38. Butler Spring local fauna, Meade County, Kansas. Pap. Surdam, R. C. and Stanley, K. O., 1979. Lacustrine Mich. Acad. Sci., Arts Letters, 50: 235-265. sedimentation during the culminating phase of Eocene 592

Lake Gosiute, Wyoming (Green River Formation). Geol. Wilson, M. V. H., 1977. Paleoecology of Eocene lacustrine Soc. Am. Bull., 90:93 110. varves at Horsefly, British Columbia. Can. J. Sci., Surdam, R. C. and Wolfbauer, C. A., 1975. Green River 14: 953-962. Formation, Wyoming: a playa-lake complex. Geol. Soc. Wilson, M. V. H., 1978. Upper Cretaceous marine Teleostei Am. Bull., 86: 335-345. from the basal Kanguk Formation, Banks Island, Swift, C., 1968. Pleistocene freshwater fishes from Ingleside Northwest Territories. Can. J. Earth Sci., 15: 1799- Pit, San Patricio County, Texas. Copeia, 1968(1): 63-69. 1807. Tasch, P., 1965. Communications theory and the fossil record Wilson, M. V. H., 1980. Eocene lake environments: depth, of invertebrates. Trans. Kans. Acad. Sci., 68(2): 322 329. and distance-from-shore variation in fish, insect, and Teller, S. and Bardack, D., 1975. New records of Late plant assemblages. Palaeogeogr., Palaeoclimatol., Pleistocene vertebrates from the southern end of Lake Palaeoecol., 32: 21-44. Michigan. Am. Midl. Nat., 94(1): 179-189. Wilson, R. L., 1968. Systematics and faunal analysis of a Teller-Marshall, S. and Bardack, D., 1978. Post-glacial Lower Pliocene vertebrate assemblage from Trego fishes from a Lake Michigan drainage in Milwaukee, County, Kansas. Contrib. Mus. Paleontol., Univ. Mich., Wisconsin. Contrib. Biol. Geol., Milw. Public Mus., 15: 22(7): 75-126. 1-19. Wolfe, J. A., 1971. Tertiary climatic fluctuations and meth- Uyeno, T., 1963. Late Pleistocene fishes of the Clear Creek ods of analysis of Tertiary floras. Palaeogeogr., Palaeo- and Ben Franklin local faunas of Texas. J. Grad. Res. climatol., Palaeoecol., 9: 27-57. Center, South. Methodist Univ., 31(3): 168-173. Zangerl, R., 1971. On the geological significance of per- Uyeno, T. and Miller, R. R., 1962. Late Pleistocene fishes fectly preserved fossils. In: E. L. Yochelson (Editor), from a Trinity River terrace. Copeia, 1962(2): 338-345. Proc. North American Paleontological Convention, Part Waldman, M., 1971. Fish from the freshwater Lower I, Extraordinary fossils. Allen Press, Lawrence, Kans., Cretaceous of Victoria, Australia, with comments on pp. 1207-1222. the paleo-environment. Spec. Pap. Palaeontol., 9:1 124. Zangerl, R. and Richardson, E. S., 1963. The paleoecologi- Weigelt, J., 1927. Resente Wirbeltierleichen und ihre cal history of two Pennsylvanian black shales. Fieldiana Pal~iobiologische Bedeutung. Weg, Leipzig, 227 pp. Geol. Mem., 4:1 352.