GAIA N° 15, LlSBOAlLISBON, DEZEMBRO/DECEMBER 1998, pp. 279-300 (ISSN: 0871 -5424)

PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY: BLENDING QUALITIES AND QUANTITIES IN THE SCIENCE OF ICHNOLOGY

Martin LOCKLEY Department of Geology, Campus Box 172, University of Colorado at Denver. P.O. Box 173364, DENVER. COLORADO, 80217-3364, USA E-mail: [email protected]

ABSTRAct: Theropod tracks are sometimes viewed as lacking much morphological variation owing to the conservatism of the theropod foot. A review of the Mesozoic track record for this group, with special reference to the western United States shows that this is not the case, and that the theropod tracks are quite diverse in comparison with the tracks of other groups. Study of large samples of well-known latest Triassic (Rhaetian) to latest Cretaceous (Maastrichtian) theropod tracks from North America, indicates that there are a variety of dis­ tinctive morphologies recognized by the ichnogenera Grallator, Kayentapus, Eubrontes, Carme/opodus, Therangospodus, Mega/osauripus, /renesauripus, Saurexallopus, Tyran­ nosauripus, and at least one other unnamed ichnotaxon. The study of theropod, mainly tri­ dactyl, tracks has benefited from the application of a wide range of different morphometric techniques, in recent years. These techniques are, however not usually applied with consis­ tency nor used as the basis for subsequent taxonomic revisions. Given the history of re­ search into theropod tracks, it is argued that most tracks have been discriminated and named using the human eye and selected techniques of morphometric analysis rather than the application of any standardized or comprehensive methodology. Although this method may sound "unscientific" it nevertheless appears to have been successful to some degree, though in other cases it has lead to the erection of ichnotaxa that are of dubious validity. Re­ cent efforts to apply objective quantitative techniques, have also failed to adhere to any standardized approach, and frequently employ morphometric methods that can hardly be described as comprehensive. Many such studies fail to pay much attention to large samples of tracks. Given the precision of the human eye and the ability of experienced ichnologists to recognize distinctive morphologies, it is argued that any attempt to quantify morphologi­ cal variation in fossil footprints, is unlikely to be successful, or widely accepted without first laying out a careful rationale regarding exactly what should be measured, and why chosen parameters are important. It is argued herein that selection of suitable morphometric para­ meters for analysis should be based primarily, though not exclusively, on the phenomeno­ logical approach of simply establishing which characters are most distinctive, and that it is impossible and often unrewarding to try and measure every variable. This approach recog­ nizes and describes diversity of form that arises from within the unity of the entire sample. By contrast the approach of pooling all data tends to artificially blur otherwise valid distinc­ tions (Le., impose unity on diversity) by creating composite assemblages that do not exist in nature. Caution is urged in the comparison oftracks from different stratigraphic sequences (series and stages), owing to growing evidence for palichnostratigraphic track zones. Simi­ larly comparison of tracks and track makers from different ages, epochs and periods is also misleading again imposing an artificial homogenization on real biostratigraphic diversity by creating potential correlations for which no geological (chronological) evidence exists. Efforts to correlate tracks ofthe same age, however, can be rewarding, though caution must be exercised in correctly determining their stratigraphic contexts.

279 artigos/papers M. LOCKLEY

INTRODUCTION purely quantitative or a purely qualitative description that satisfies general scientific and taxonomic con­ How WE OBSERVE: MORPHOLOGICAL QUALITIES vention). The extentto which the description that dif­ AND QUANTITIES ferentiates two morphotypes is applicable to other theropod tracks should also be considered, but can Scientists are conditioned, to some extent, to be­ only be standardized if the assumption is made that lieve that in order to be scientific in our approach to all other tracks share similar features. This assump­ organisms, we must "objectively" measure and tion is likely to be valid in most, but notall, cases. As I quantify their component parts, and understand the hope to show, vertebrate ichnology provides a good mechanisms by which they function. As a result opportunity to review our methodologies, and I have modern biology has sometimes been regarded as chosen the subject of variation in theropod tracks as unduly mechanistic, and inclined to ignore the attrib­ a case study. utes of the entire organism and the dynamic pro­ cesses of function (cf. SCHAD, 1971; LOCKLEY, Theropod footprints are relatively simple mor­ 1999a). For example we may study the genetic ma­ phologically, consisting of the impressions of three ke up of an organism without really looking at the or­ digits (11,111 and IV) in most cases, though in some ganism itself as a complex system (GOODWIN, footprints the impression of a fourth digit (the hallux 1994). Morphometric analysis also reduces a com­ or digit I) is present (Fig. 1). Other vertebrates in­ plex array of attributes of an actual organism to a cluding ornithopod dinosaurs, stegosaurs, pro­ matrix of numbers that are no longer in biological sauropods and birds may also leave similar tridactyl context. The utility of such quantitative data, is to a and tetradactyl tracks. For the purposes of conven­ significant extent dependent on the assumptions ience I will refrain from listing all the other verte­ made by the researcher collecting and analyzing the brates from crocodiles to ceratopsians that can data. For example if the integrity of the sample is leave tetradactyl tracks. But I will ask what would questionable (e. g., not representative of a popula­ seem to be a logical question. In order to narrow the tion in either a biological or statistical sense), or the field of study to theropod tracks, or at least to tracks measurements selected, are not meaningful with re­ that have a good probability of being of theropodan spect to the organism's growth or anatomy, or if affinity, how do we distinguish them from the foot­ questionable statistical tests and interpretations are prints of other vertebrate groups, that in a strict applied during synthesis, the utility of both the data quantitative sense (number of digit impressions) and the resultant interpretations may be compro­ have a comparable morphology? Have quantitative mised. analyses and morphometric studies been under­ taken to discriminate major track groups in the first This perspective also applies, quite forcibly to place? The answer is almost always no. Except in a paleontology, where it is not possible to examine the few cases (e.g., OLSEN, 1980; MORATALLA, SANZ & complete organism or observe itfunctioning as a dy­ JIMENEZ, 1988; DEMATHIEU, 1990; WEEMS, 1992; namic living organism. Thus the temptation is per­ FARLOW & LOCKLEY, 1993; FARLOW & CHAPMAN, haps even greater, than in Biology, to measure and 1997), little in the way of quantitative analysis has quantify the limited amount of material available to been presented. It is worth noting here that the ana­ squeeze out all available "data". Because paleontol­ lytical techniques employed by these authors are far ogy is concerned with evolutionary relationships, from being standardized, and some do not even ap­ cladistic analysis helps to standardize, and objectify proach the quantitative study of tracks through the (though not quantify in a strict sense) the characters footprints themselves, but rather deal with the mor­ that are selected, defined and compared. But again, phology of foot bones of potential trackmakers. the subjectivity that creeps into the selection and definition process, and the choice of analytic tech­ None of these studies have yet made broad con­ niques, colors the eventual "results" and interpreta­ tributions to the differentiation of a wide range of dis­ tions. This is not to say that any ofthe analytical and tinct morphotypes. On the contrary they have quantitative techniques in current use should be tended to suggest that tridactyl tracks are hard to dif­ abandoned, unless they are shown to be ineffective ferentiate. But this may be in part because they se­ or misleading. Rather, I wish only to remind practitio­ lected ichnogenera which show limited mor­ ners that we must periodically review our methodol­ phological variation relative to one another. As I ogy in the context of our objectives. hope to show the inability to achieve striking suc­ cess in distinguishing different tridactyl and tetra­ For example if everyone agrees that two thero­ dactyl track types, is not necessarily a function of pod track morphologies are obviously or even subtly morphological homogeneity or conservatism in the different, then the main objective should be to con­ feet of theropods and other vertebrates, though this cisely describe this difference by appropriate may be a partial explanation. Particularly in the case means, whether quantitative, qualitative or both. (In of certain samples such as those from the classic this regard it is practically impossible to have a

280 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

width tirety, (i.e., the sum of morphological features, such as size, slenderness, digit divarication and relative claw length, pad definition, orientation relative to track­ way configuration, recurrent stratigraphic position, \ recurrent facies relationship) have "qualities" that, at c: pad -.: least in most cases, make different morphotypes quite distinguishable to most ichnological observ­ \ eo o ers. Another quality might be having a look similar to \ ... that of the corresponding foot skeleton. The differen­ \ right tiation is done with the human eye before any analy­ sis, and sometimes no thorough analysis follows. w Here we are reminded of Medawar's classic essay c "Is the scientific paper a fraud" (MEDAWAR, 1996) in ....a: which he argues that in most cases scientists essen- - en ~ tially arrive intuitively or serendipitously at an opin­ I ion about their objects of study (in this case track I morphology) before they retroactively present data and rational arguments that suggest a logical pro­ I~;;; .. cess of scientific explanation and deduction . ... " 0 ItIJ~ ii ..• This perspective is not to suggest that tracks or I any other objects of study lack real qualities of mor­ phology and geological context that must be consid­ ered in any thorough analysis. In reality however, as @ it is impossible to measure "everything" ichnologists . select only some of these qualities to differentiate distinct track types and apply ichnotaxonomic TRACKWAY names. Some times this is done judiciously, and WIDTH sometimes it is not, leading to recognition ofa further intangible and subjective "quality" that we might call the skill of the ichnologist. This quality of skill can Fig. 1 - Standard track and trackway measurement pa­ only be judged orevaluated in the fullness oftime, by rameters. the extent to which other ichnologists accept the va­ lidity of work. In the case of vertebrate ichnology, however, it is known that, until recently, there has Lower Jurassic of the Connecticut Valley, it may be been little serious attempt to evaluate classic work, in part the result of comparing tracks that are very so the longevity of an ichnotaxon must also be evalu­ similar, partly as a result of uncertainty about the ori­ ated in reference to the frequency with which the gin and integrity of samples from localities that are footprints in question have been the subject of seri­ no longer accessible, and partly the result of not ous study. making the right measurements and observations in the first place. In a recent study, using the Schad ian perspective (LOCKLEY, 1999a, b), it has been suggested that the­ The next question of many logical questions to re are recognizable morphological and evolutionary ask is: given the lack of standardized quantitative gradients associated with dinosaur foot morphol­ study are we at a loss to differentiate footprints made ogy, that reflect overall anatomy. For example thero­ by these major vertebrate groups? The answer is pod tracks are narrower than ornithopod tracks (less not, as one might expect, yes in most cases, butgen­ digit divarication) because all saurischians have erally no. Various compilations (e .g. HAUBOLD, narrow feet than ornithischians of equivalent size. 1984; LEONARDI, 1987; THULBORN, 1990; LOCKLEY, This narrow-wide polarity, in saurischians and or­ 1991) present catalogues of differentiated track nithischian respectively, is also reflected in the types that, to the best of our knowledge, are ade­ shape ofunguals (i.e., laterally or parasagitally com­ quately differentiated in the majority of cases. How pressed rather then dorso-ventrally compressed, can this be, however? If we have not used quantita­ especially in larger forms). It is also evidentthat in all tive means or standardized analytical techniques to dinosaur clades there is a tendency towards greater differentiate footprints, how can we possibly know fleshiness of the foot as size increases. This trend which morphological groups they fall into and which can be explained as a function of size, but it is also a groups of trackmakers were responsible for making refection of an intrinsic morphodynamic pattern them? The answer is simply that different track mor­ (LOCKLEY, 1999a, b). This simple observation goes photypes, when viewed "holistically" or in their en-

281 M. LOCKLEY

a long way to explaining why smaller more primitive phers of science such as Bortoft, himself a mathe­ theropods from the Jurassic often have less fleshy matician, verge on pointing out that the "mathemati­ feet with discrete digital pads that reflect the shape zation" of natural objects is a type of metaphysical . of the thinly covered foot skeleton, and why larger abstraction, that can lead to our overlooking other more derived theropods from the Cretaceous often essential ·qualities of the type oullined in the pre­ have fleshy feetthat hide the foot skeleton anatomy. vious section. All this does not mean that we can not, or should not make prolonged and detailed observa­ WHAT IS IMPORTANT AND How TO LOOK OUT tion of the quantities (and qualities) of the objects of FOR IT our study, to gain a deeper understanding. However it does suggest that different criteria, and different The next of many pertinent questions is how powers of discrimination, may be required to differ­ many qualities and features, could and should be entiate different objects (tracks) in differentcases. quantified in order to serve the objective of different i­ ating different track types. There may be more than In the sections that follow I first review various one acceptable answer to this question, including methods that have been used to discriminate thero­ the obvious, but quite different, answers "as many pod tracks, and, secondly, I select a variety of re­ as necessary" or "as many as possible." These an­ cently studied theropod ichnogenera, and attemptto swers imply that the study will proceed until differ­ show the features (qualities and quantities) which ences are found, or until all possible means of make them distinct taxonomic entities. I have taken differentiation are exhausted. In the former case the the easy, but legitimate route of focusing attention study ceases, when the preconceived (subjective, mainly on tracks that have obvious and easily identi­ but often valid) goal of differentiation is achieved. In fied characteristics. I have also focused attention on the latter case, the study continues until no further samples with which I am familiar, and for which sub­ methods of analysis can be conceived. I would ar­ stantial data sets exist. In this regard a concrete gue here that, since new techniques are regularly in­ qualitative basis exists for further analysis (i.e. I sta­ troduced in science, the latter eventuality is te explicitly what are the morphological features that theoretically impossible and has never been ap­ are perceived to set different tracks apart from each proached in the study of tracks where new analytical other). Such observations might help pave the way methods are being added to the forty or so applied in towards compiling a discrete list of characters so studies to date. that we might attempt a cladistic analysis. Such ob­ servations also allow us to consider the potential for It may, in some cases, only be necessary to de­ quantitative approaches that more carefully decide scribe a few distinctive features which indicate that a which features should be measured in order to bet­ particular morphotype is different from any others ter understand real morphological characteristics that are known. Such would appear to be the case, that make the tracks distinctive in the first place. for example, with such unique footprints as Tyran­ From here we can consider what features warrant nosauripus and Saurexallopus, discussed below. further study. Whereas in other cases the analysis and description of multiple qualities, and quantities, may still prove EXAMPLES OF QUANTITIES THAT HELP inadequate to the task of providing clear differentia­ DEFINE TRACK MORPHOLOGY tion. This may be the case with the Grallator, Anchi­ sauripus and Eubrontes plexus (GAE), which, ac­ Ichnologists routinely measure such features as cording to OLSEN (1980), all form part of a complex track length, track width, the length of individual digit morphological continuum. But, as I argue below, impressions, the divarication between digit impres­ these tracks have, as yet, only been subjected to lim­ sions II and IV, the step and stride length and pace ited, and demonstrably disparate, analyses. Here, I angulation (e.g., BAIRD, 1957; LEONARDI, 1987; would venture a simple observation. Obvious mor­ THULBORN, 1990; LOCKLEY , 1991: see Figure 1 phological uniqueness is a characteristic quality of herein). In some cases other parameters such as in­ Tyrannosauripus and Saurexallopus relative to ward or outward rotation of the long axis of the foot­ other theropod tracks, and morphological similarity print is measured (LEONARDI, 1987). It is common to is a characteristic quality that emerges from a com­ take the ratio of foot length to foot width as an indica­ parison of the three ichnogenera in the GAE plexus. tion of whether the foot is "elongate" or "transverse." They are, at least on one level, what we see on initial Such qualities of shape may be used to interpret the inspection. It is importantto pay attention tothis phe­ affinity of the trackmaker. For example theropod nomenological perspective, in order to avoid the trap tracks are characteristically elongate. of departing from the realm of concrete observation Some authors such as OLSEN (1980: fig. 21) and too far into the realm of quantitative abstraction WEEMS (1992: fig. 2) have suggested using the len­ (SCHAD, 1971, BORTOFT, 1996). Atthe risk ofcriticiz­ gth to which digit III extends anteriorly, beyond a line ing the quantitative approach too strongly, philoso- drawn transversely from the tip of digit II to the tip of

282 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

r---- I I I I I I 1 :;:1 1 I I 1 I I L __ c (fI-tel/fw

Fig. 2 - Track measurement parameters employed by OLSEN (1980) and WEEMS (1992).

digit IV (Fig. 2), as measure of shape that will help descriptions. Recently however, OLSEN, SMITH & discriminate tracks. It is already known that this McDoNALD (1998) indicated that, based on type value is related to some degree to digit divarication, specimens, Grallator, Anchisauripus and Eubrontes which may be influenced by preservational differ­ can again be regarded as distinct ichnogenera! ences between tracks. According to OLSEN (1980: Thus it seems we are back to square one. 370), in two paragraphs of discussion, specimens Here some further methodological observations "typical of Grallator, Anchisauripus and Eubrontes are pertinent. Olsen mentions no morphometric pa­ show differences in proportions which probably re­ rameters of theropod tracks other than the great fiect real differences in foot structure (BAIRD , 1957); variation in divarication (15-45°) between digits II the main factorresponsible forthis, however, is prin­ and IV, for use in discriminating ichnotaxa! So we do cipally the relative length of digit III. A reasonable not know how length and width ("elongateness") measure of this factor is the projection of digit II past varies, how digit divarication varies between digits II and IV." Olsen goes on to say thatifthis measure is II , III and IV, for each named ichnotaxon, how step, graphed it becomes clear that the "shape of the pes stride and pace angulation, or rotation of the foot changes continuously with size," which might "be might vary independent of the "projection of digit II expected of tracks of individuals of different ages pastil and IV." This is to say nothing of detailed mor­ perhaps representing a single dinosaur species. It is phological features of individual, digits, pads, hallux therefore reasonable to synonymize the junior orhypex to name justa few potentially importantfea­ names Eubrontes and Anchisauripuswith the senior tures. In short, Olsen's two paragraphs on continu­ name Grallator." But, is it reasonable to synonymize ous variation in a single variable do not provide an these three ichnogenera on the basis of what ap­ adequate model for comprehensive analysis in the­ pears to be continuous variation in a single variable? ropod footprints. Contrary to what one might expect, from his argu­ ment of similarity Olsen concludes that he does not WEEMS (1992) adopted the idea of measuring believe that there was only one species of theropod the "projection of digit II past II and IV" and labeled dinosaur trackmaker, and that further study is re­ this parameter "toe extension" (te). He then meas­ quired . Despite this conclusion some authors (e.g. , ured foot length (fl) and width (tw) for "( ... ) 7 ichno­ GIERLINSKI, 1991; GIERLINSKI &AHLBERG, 1994) fol­ genera and 23 ichnospecies of tridactyl saurischian low OLSEN (1980) in using the double-barreled ich­ dinosaur tracks previously described in the litera­ nogenus names of Grallator (Grallator), Grallator ture." (WEEMS, 1992: 113), and plotted (fl-te)/tw (Anchisauripus) and Grallator (Eubrontes) in formal against te/tw. In short rather than simply plotting the

283 M. LOCKLEY

extension of digit II past II and IVagainstthe remain­ phalange 1 ( = metatarsophalangeal joint) except der of foot length, Weems was plotting these two di­ perhaps in digit IV (BAIRD, 1957; THULBORN, 1990; mensions relative to foot width. In other words he FARLOW & LOCKLEY, 1993: 123). Instead, especially defined the shape of the two triangles (posterior and in digits II and III, the posterior pad impressions rep­ anterior) defined by the heel and the extremities of resent the articulations between phalanges 1and 2. digits II, III and IV. In adopting this procedure Weems FARLOW & LOCKLEY (1993: fig. 2) also plotted the ra­ was essentially supporting Olsen's proposal that tio of length of phalanx 111211VI against 1131112 (Fig. 3) theropod tracks can be differentiated on the single on the assumption that theropods usually have a parameter of toe extension, in this case considered relatively longer phalanx 112 (compared with the in relation to foot shape. Again it can be argued that lengths of 112 and IV1) than do prosauropods and or­ such an approach is not comprehensive. Nonethe­ nithischians. less, Weems arrived at some specific conclusions Despite the intricacies of this approach, Farlow's as follows. He inferred thatthe 7 ichnogenera and 23 main objectives (SMITH & FARLOW, 1996; Farlow, ichnospecies examined could be reduced to three pers. comm.) have been to differentiate theropod ichnogenera (Grallator, Eubrontes and Kayenta­ tracks from those of other tridactyl (or tetradactyl) pus) containing only nine ichnospecies (distributed trackmakers. His methods, however, have not pro­ 5, 2 and 2 respectively in relation to the three ichno­ vided a foolproof method of differentiation, as these genera). He showed that Grallator tracks are more particular "osteometric ratios" do not clearly "dis­ elongate than Eubrontes tracks, suggesting that the criminate theropod from tridactyl ornithischian latter, larger trackmaker had to distribute its weight tracks" (FARLOW & LOCKLEY, 1993: 130). In short across all three toes. (LOCKLEY, 1999a: fig . 4 .3 Farlow's objectives have generally not, as yet, been reached a similar conclusion about overall narrow­ to differentiate different ichnotaxa, on the basis of ness and width, based on recognition of inherent sample by sample comparisons, as done by morphodynamic gradients in the saurischian/thero­ Weems, or even by Olsen. Farlow (SMITH & FAR­ pod foot). LOW, 1996, and personal communication) has a par­ Apparently the utility of Weems study, building on tial, and intriguing, ulterior motive of trying to estab­ Olsen's "ideas," is debatable, even though it poten­ lish the extent to which ichnotaxa correspond to tially provides ichnologists with some explicit tools osteological taxa. I shall return to this topic later, but for comparing different track types. Based on other note here that this goal does not necessarily justify approaches Weem's method is incomplete, how­ imposing osteologically derived preconceptions or ever, it at least resulted in some explicit and signifi­ rules on the morphological description of tracks. cant ichnotaxonomic inferences and conclusions. Track morphology should first be described for what Here it is necessary to add that in a recent restudy of it is, then later interpreted for what it represents. the Grallator, Anchisauripus and Eubrontes type Even though it is pleasing to match tracks with track- material OLSEN, SMITH & McDoNALD (1998) make no reference to WEEMS ' SUPPRESSION OF Anchi­ sauripus and his recognition of Kayentapus. At the risk of chiding my colleagues, it is poor scientific practice to ignore such work. If one disagrees with the conclusions of another one should state why and attempt to offer alternatives. FARLOW & LOCKLEY (1993) suggested another approach to the morphometric analysis of theropod tracks. Based largely on unpublished research by Farlow (see also SMITH & FARLOW, 1996, and FAR­ LOW & CHAPMAN, 1997) the length of digit impres­ sions II and IV relative to the impression of digit III was analyzed by plotting digit III length/digitll length against digit IIllength/digit IV length. The fields de­ rived from this method can be compared with the fields derived from plotting the same parameters for actual foot bones. A potential drawback is that one must infer the extent to which the sets of phalanges (113 and 112; 1114, 1113 and 1112; IV5, IV4, IV3, IV2 and IV1) actually correspond to the digit pad impres­ sions seen in footprints. For example, the posterior Fig . 3 - Pad measurements (after FARLOW & LOCKLEY, pad does not always represent the posterior end of 1993). Scale bar = 10 em.

284 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

makers, and develop an ichnotaxonomy that re­ be derived from deep tracks even when the trace flects this correspondence of the body and trace appears to suggest a different angle. fossil records, such an achievement is not always It has also become clear that in some tridactyl, or possible. functionally tridactyl species, footprints clearly show that there is a' marked asymmetry between the "pro­ EXAMPLES OF QUALITIES OBSERVED IN ximal" or posterior depth of the hypex (including sha­ TRACKS. pe and angle of indentation) on either side of the im­ Theropod tracks, such as Grallator, from the pression of digit III, (i.e., the difference in the hy­ Lower Jurassic of eastern North America, often re­ pexes between digits II and III and between digits III veal well-defined digital pads, separated by obvious and IV: see Fig. 4). Such "qualities" or features might creases. Recent studies, however, have revealed be described as differential anterior or posterior hy­ that some tracks lack such distinctive or differenti­ pex configuration. The location or configuration of ated pads (e.g., Therangospodus: LOCKLEY , MEYER the hypex is a primary morphological feature that - & MORATALLA, 1998). Thus the lack of differentiated tells us about the articulation of the foot. Therefore, pads, which is not attributed to poor preservation, is there may be no direct relationship between the nu­ a primary morphological feature or "quality" that merical value of the angle and the location of said does not lend itself to immediate quantification. Con­ angle in some cases, though in other cases the dif­ sequently it is not possible to measure pad dimen­ ferential values of the angles may be significant. sions for comparison with other track samples. Si­ Such features can be measured, using multivariate milarly some theropods appear to have very wide techniques such as have been applied by undifferentiated pads on digit II and narrow more dif­ MORATALLA, SANZ & JIMENEZ (1988). ferentiated pads on digit IV. Similarly the presence or Other significant morphological features orquali­ absence of claw impressions, though preservation­ ties that have been observed are the extent to which ally controlled to some degree, may also be the re­ individual digit impressions taper distally, are paral­ sult of a primary morphological characteristic, that lel sided or fusiform. Again such features can be relates to size of claw, the extent andlor angle of pro­ measured, using multivariate techniques, and tend trusion from fleshy portions of the foot, etc. Recently to give results that indicate that digit width and shape GATESEY et al. (1999) have show that the true "mor­ are very significant factors in "explaining sample phological" configuration (or angle) of the hallux can variability" (MORATALLA, SANZ & JIMENEZ, 1988).

III

10=] IV II IV

10=]

Fig. 4 - The quality of "indented ness" highlighted by a comparison of the Upper Cretaceous theropod track Saurexa/­ lopus from Wyoming (right) with a newly-discovered, unnamed, Upper Cretaceous theropod track from South Dakota (left), reveals opposite patterns of indentedness in between digits II, III and IV. Such patterns are primary morphological features that reveals details of the foot morphology of the trackmakers.

285 M. LOCKLEY

Such observations need to be contrasted with the in­ SHAPE ANALYSIS AND LANDMARKS terpretations of OLSEN (1980) and WEEMS (1992), Various forms of shape analysis can be applied to which take no account of digit shape. As Indicated In tridactyl tracks as proposed by MORATALLA (1993) the following section, attempts to quantify such fea­ and RASSKIN-GUTMAN et al. (1996, 1997). Shape tures or qualities can easily result in the recognition analyses have various advantages and disadvan­ of dozens of characters that can potentially be mea­ tages. They use a series of landmark POints to define sured). It is worth noting that often qualities such as an outline for a track (Fig. 6), and employ "high-Ievel fusiform shape, tapering, narrowness, or breadth mathematical, geometric and statistical approaches are easily discernable with the naked eye: prior to to quantify the shapes of objects." Consequently being subjected to measurement. In quantitative a­ such analyses have the potential to dlscnmlnate dif­ nalysis such qualities can not be expressed on a ferent track morphologies, including such features simple linear scale, and so are often expressed as as digit width, tapering, divarication angles etc., ra­ ratios or dimensionless multivariate numbers, or therthan only simple parameters such as length and fields in morphospace. In this regard they become width. The disadvantages of the method, as pres­ just as abstract, conceptually, as the initial qualita­ ently employed, are that they do not allow for analy­ tive perceptions (such as fusiform shape) that they sis of morphological features (such as pads, skin seek to describe. It is a common mistake to think that texture etc.) found within the outline of a track, the quality (e.g. fusiform shape) is not real until it is though presumably such morphologies can be ana­ verified by quantitative analysis, when in fact it was lysed. For the practicioner to fully appreciate what observed and recognized qualitatively from the out­ they are analyzing and discriminating, application of set. Such epistemological considerations stress the the method also requires an advanced understand­ need to pay more attention conscious attention to ing of"high level" mathematics, geometry and statis­ our initial empirical observations. tics, and suitable facilities for undertaking analyses. MULTIVARIATE ANALYSIS THE CLADISTIC APPROACH MORATALLA SANZ & JIMENEZ (1988) carried out Cladistic analysis of tracks has only been men­ a multivariate ~nalysis of 66 Lower Cretaceous tri­ tioned and attempted in a perfunctory way (OLSEN & dactyl footprints from "diverse geographical origins" BAIRD, 1986; OLSEN , SMITH & McDONALD, 1998; in order to determine if theropod and ornithopod PITTMAN, 1992). Theoretically the problem with this tracks could be differentiated. The features meas­ approach is that only a limited number of useful non­ ured were divided into two categories, non-metric metric characters are available for comparison. For and metric. Non-metric parameters are akin to quali­ example, in the studies cited, only about 10-2.0 per­ ties that do not lend themselves to measurement, cent of the total list of potentially useful metric and such as presence or absence of claws, the presence non metric characters or measures used in the study or absence of pads, the presence or absence of in­ of tridactyl tracks, have been shown to be useful in terdigital webbing (a dubious feature that, in most if attempting a cladistic analysis. In practical terms, a not all cases, can be explained as an extra morpho­ thorough cladistic analysis has never been attem­ logical or preservational feature), and heel shape pted based solely on track characteristics. The which is "difficultto systematize." Metric parameters temptation to fit tracks onto preexisting cladograms were divided into linear parameters: length, width, that are based on osteological evidence (ct., OLSEN total digit lengths (LlI, LIII, LlV), basal digital lengths & BAIRD, 1986; PITTMAN, 1992) is just anotherwayof (BL2, BL3, BL4), basal digit width (WBII , WBIII, trying to match tracks with trackmakers, and appar­ WBIV), middle digital width (WMII, WMIII, WMIV) ently has no obvious utility or potenllal for the mor­ and heel-interdigital distance (K, M for distance from phological comparison and differentiation of the heel to medial and lateral hypex (Fig. 5). Metric, an­ tracks themselves. gular parameters: interdigital angles II-II and II-IV (alpha and beta respectively). Arising from these SYNTHESIS OF APPROACHES various parameters two categories of "morphomet­ ric ratios" were obtained: those "related to dimen­ A critical analysis of previous work indicates that sions of the whole footprint. LlW, LlK and LIM" and almost everyone who has attempted to measure those "related to digital morphology. BL2/WMII, theropodltridactyl track morphology has employed a BL3/WMIII, BL4/WMIV; Lli/WBII , LIII/WBIII, different methodology. Such a state of affairs could LlV/WBIV." (MORATALLA, SANZ & JIMENEZ, 1988: be viewed uncharitably as indicating a lack of stan­ 398). This constitutes a total of 32 parameters (four dardization in vertebrate ichnological studies. How­ nonmetric, seventeen metric linear, two metric an­ ever the science is immature, and most of these gular, and nine dimensionless ratios: see Figure 5. tech~iques have been applied only in the last five to ten years. It is also evident that different techniques, meet with different degrees of success. ThiS returns

286 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

w

L

Fig. 5 - Parameters used in multivariate analysis (after MORATALLA, SANZ & JIMENEZ, 1988).

us to our previous question of how many parameters o the length of digit impressions II and IV relative should be measured, and how many is it actually to the impression of digit III and the morphospace necessary to measure. Again from a philosophical field derived from the ratio of digit III length/digit II point of view, it is impossible to legislate absolutely length plotted against digiti II length/digit IV length. what should be measured, for once a list is agreed o the ratio of length of pads corresponding to upon, even by the wisest committee of ichnologists, phalanx 1112/IVI plotted against 1131112 it leaves no room for introduction of new parameters or abandonment of old parameters that may be presence or absence of claw impressions shown to be redundant. A balance must be struck o presence or absence of pad impressions between adherence to methodologies shown to be effective by informed practicioners of ichnology, and o digit shape (degree oftapering, parallel-sided- the introduction of novel or specific techniques suit­ ness, or fusiformness) able in particular circumstances. total digit lengths (LlI, LIII, LlV) Based on the above discussion of parameters basal digital lengths (BL2, BL3, BL4) used in the description and differentiation of thero­ pod, and other tridactyl-tetradactyl tracks, we can basal digit width (WBII, WBIII, WBIV) compile the following list of potentially useful meas­ middle digital width (WMII, WMIII, WMIV) urements (quantities) and features (qualities): heel-interdigital distance (K, M for distance o Foot length (fl) and foot width (fw) and the deri­ from heel to medial and lateral hypex) ved ratio, or measure of shape: fl/fw. o interdigital angles II-II and II-IV (alpha and beta o Toe extension (te) = the projection of digit II respectively) past II and IV and the derived ratios (fl-te)/fw and te/fw which describe the shape of the anterior and o indented ness in relation to location, not value, posterior triangles of interdigital angles

287 M. LOCKLEY

tiple parameters. Such an observation could be in­ '1/ terpreted to mean that previous taxonomic conclu­ sions were premature, and that the material in ques­ tion should be restudied using more sophisticated techniques. This however is not a fair assessment since historically, the newer synthetic techniques were not available. We should not wish away, ignore

12 or underestimate archaic ichnotaxonomy, we must ,v amend it. It is precisely those who introduce new and potentially powerful techniques of analysis who 1:5 18 should test them out on known samples, and derive 17 taxonomic conclusions. In this regard the most com­ prehensive analysis so far proposed (MORATALLA, SANZ& JIMENEZ, 1988) essentially demonstrates its potential to differentiate tracks without drawing any taxonomic conclusions. One possible conclusion is that ichnologists are afraid of, or at least very wary of ichnotaxonomic investigations. The reasons usually given for this are that archaic taxonomic studies led to over-splitting, or that variation is hard to define owing to extra morphologic, preservational influ­ ences. I contend thatthis only part ofthe story. In the late •• twentieth century, the rise of morphometric and cla­ distic consciousness has made paleontologists Fig. 6 - Describing track outlines using "landmarks" aware of the complexity of inter- and intra-species (shape analysis), after MORATALLA (1993). variation. This make labeling of variability or the sub­ division of morphological continua into discrete classes (taxa) conceptually problematic. Butthere is • morphometric ratios related to dimensions of a significant irony here if new methods make it har­ the whole footprint, LlK and LIM der to differentiate morphologies and species. Are we scared away from traditional ta xonomy by the so­ • morphometric ratios related to digital morpho­ phistication of our new methods and current cladistic logy. BL2IWMII, BL3IWMIII, BL4IWMIV; LlIIWBII, fashions. Are we really afraid to amend and revise LlIIIWBIII, LlVIWBIV." (op cit p, 398) the work of HITCHCOCK (1858) and LULL (1953) track outline as measured by landmark analy- when we in fact have much better analytic tools at sis our disposal than they ever did. What good are new techniques if we don't use them to arrive at taxo­ cladistic analysis of tracks nomic conclusions. Conceptually it is hypocritical to Clearly there are many different (about40) quan­ arrive at the abstract conclusion that new techni­ titative, qualitative and synthetic measures that can ques reveal old techniques to be inadequate, if the be applied to track description and differentiation. new techniques are not clearly shown to produce The quantitative approach is based to some extent significant and concrete results. In the absence of on the assumption that there is so much overlapping testing of the type proposed by WEEMS (1992) the variation in tridactyl (and related tetradactyl) mor­ argument that new, more analytical, more quantita­ phology, that only an objective quantitative analysis tive, more powerful techniques are better, has a hol­ will suffice to distinguish different morphotypes. low ring. However, it is also clear that non-metric qualities and In the absence of consensus on which new meth­ ratios that define shape are also important in helping ods are most desirable and effective for advancing differentiate tracks. Similarly multivariate analysis our taxonomic understanding, and in the absence of that define dimensionless morpho space fields are taxonomic case studies arising from the application also helpful in defining morphospace variation even of these new methodologies, it is also legitimate to if their subsequent interpretation is complicated. conclude that taxonomic studies will continue in the Ironically, until now most taxonomic studies have vein of traditional or tried and true approaches. My arisen from simple qualitative studies and quantita­ sense is that we are testing out new methods without tive studies thatfocus on a few morphological meas­ applying them either extensively or comprehensi­ ures rather than those that involve analysis of mul- vely, and that until a newconsensus is reached, radi-

288 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

cal changes can not be expected overnight. No indi­ Lower Jurassic of Europe (e.g., LAPPARENT & MON­ vidual authority can hope to call a successful mo­ TENAT, 1967; DEMATHIEU, 1990, 1993; DEMATHIEU & ratorium on the scientific status quo, without good SCIAU, 1995; GIERLINSKI , 1991; GIERLINSKI & reason and the consent of others in the field. For pre­ AHLBERG , 1994), in many cases these authors have ciselythe reasons expressed above, my philosophi­ introduced different specific names. In some cases cal ruminations are framed within the context of a there is no clear reason, based on morphological cri­ review, in order to achieve the practical end of an teria for the introduction of new species names, and overview or synthesis, in which all approaches are it is therefore reasonable to regard some of these considered. In such a context the question of the in­ ichnospecies as a type of "geographic label" (LOCK­ dividual and collective utility of these approaches LEY & MEYER, 1999). and the desired objectives of such applications can For example, in their study of the Lower Jurassic be brought into the clear light of day. tracks of Le Veillon (LAPPARENT & MONTENAT, 1967) named nine new ichnospecies. Four of these, were THE THEROPOD TRACK RECORD attributed to Grallatorand Eubrontes, but three were The theropod track record is here considered assigned to the new ichnogenera Saltopoides, Ana­ with special reference to the western United States topus and Talmontopus. Their smallest ichnospe­ (LOCKLEY & HUNT, 1995). It is recognized from the cies, Grallator oloneusis, was described as being outset that the track record for the Lower Jurassic is very similar to G. tenuis and G. gracilis from Con­ the most complex and problematic from an ichnotax­ necticut. Similarly Grallator variabilis, was de­ onomic viewpoint, owing the long historyofintroduc­ scribed as being comparable or "homologous" to G. tion of new names. For this reason, the Lower cuneatus from Connecticut. Their Grallator maxi­ Jurassic theropod track record must be discussed in mus, was casually compared with Anchisauripus a broader, global context. As demonstrated by the from Connecticut. The largest tridactyl footprints as­ utility of Middle and Upper Jurassic tracks for pal i­ signed to Eubrontes veilloneusis was described as chostratigraphic correlations other track assem­ being "most close" to E. giganteusfrom North Amer­ blages are also best considered in their global ica. When we closely examine the newly named ich­ context. nogenera Saltopoides and Anatopus both fall within the size range and morphological range of Grallator UPPER TRIASSI C variabilis and neither warrant being placed in a se­ parate ichnogenus (LOCKLEY & MEYER, 1999). As The Upper Triassic track record is dominated by shown below, Talmontopus may also be compared small theropod tracks attributable to ichnogenus with various ichnospecies of Grallator. Grallator. In many regions the dominant track type is a small "ichnospecies" no more than 6-7 cm in len­ WELLES (1971) in a study of Lower Jurassic gth, with narrow trackways (LOCKLEY & HUNT, tracks from Arizona, introduced the names Kayenta­ 1995). Largertracks, 12-15 cm long, and also attrib­ pus and Dilophosauripus for large tracks that might utable to Grallator, are found at some sites, but in be synonymous with Grallator, Anchisauripus lesser abundance, and a few large tridactyl forms, andlor Eubrontes (LOCKLEY, 1986; LOCKLEY & that have not been named are also known. Abun­ HUNT, 1995). WEEMS (1992) however considers that dant, small and ubiquitous Grallator is therefore Kayentapus should be recognized as a distinctive characteristic of the earliest expression of the thero­ ichnotaxon. IRBY (1995) the only author, since pod track record, and can be regarded as an impor­ Welles, to formally study Lower Jurassic tridactyl tant ancestral morphotype. tracks from this region, named Grallator (Anchisau­ ripus) madseni, apparently distinguishing a new ich­ LOWER JURASSIC nospecies from known taxa on the basis of a "pos­ terolateral marking" behind the impression of digit As noted above, much of the debate about the IV. The tracks she described, however, are pro- ba­ nomenclature of Lower Jurassic theropod tracks re­ bly not the same as those assigned to Kayentapus. volves around whether the names Grallator, Anchi­ Herwork however serves to confirm the presence of sauripus and Eubrontes, are valid as distinct tracks that are attributable to Grallatorin this region ichnogenera. These well known ichnogenera were (cf. LOCKLEY & HUNT, 1995). introduced by HITCHCOCK (1858, and earlier publi­ cations) and accepted by LULL (1953 and earlier It is outside the scope of this paper to discuss publications), leading to more than a century of con­ whether "medium sized" Kayentapus is distinctfrom sistent usage. It is clear however that this consistent small Grallator and large Eubrontes, as suggested usage has been restricted to the use of ichnogenus, by WEEMS (1992), so I tentatively accept his conclu­ not ichnospecies names. For example while Gralla­ sion pending further study, and note that OLSEN et al. tor and Eubrontes have been used as ichnogenus (1999) have also avoided discussion of the topic. categories in the description of similar tracks in the Based on field experience in the western United

289 M. LOCKLEY

States, however, and a survey of literature on Lower Grallator and Eubrontes from the same localities Jurassic tracks, it appears that there are Kayenta­ (LAPPARENT & MONTE NAT, 1967), nor is it clearthatit pus-like tracks intermediate in size and shape be­ is suitably-preserved with respect to details of indi­ tween diminutive slender morphologies such as vidual pad impressions to serve as a type for what Grallator cursorius, Grallator gracilis and Grallator appears to be a very widely distributed morphotype. tenuis (HITCHCOCK, 1858; LULL, 1953) and large ro­ bust forms such as Eubrontes giganteus. These MIDDLE JURASSIC forms can be described as being relatively large, Fortunately, because there is little tradition of with less tapering digits and greater digit divarication work on Middle Jurassic tridactyl tracks, there has angles (especially between III and IV) than the afo­ not been the proliferation of names encountered in rementioned Grallator ichnospecies, but more gra­ studies relating to Lower Jurassic footprints. In fact, cile than robust Eubrontes. It has already been sug­ to date, there has only been one study in which a gested that such tracks may be similar to Schizo­ large sample of tracks from a single stratigraphic grallator from China (ZHEN et a/., 1989; LOCKLEY & unit has been described and named in detail (LOCK­ HUNT, 1995). Examples of this track type, from the LEY et a/. , 1998b). In this study the new ichnogenus western United States (LOCKLEY et a/., 1998a), Carmelopoduswas identified on the basis of the dis­ Europe and Asia are shown for comparison (Fig. 7). tinctive phalangeal pad formula 2-3-3 correspond­ Among those tracks from the Lower Jurassic, ing to digits II, III and IV. Such a morphology that could be considered Kayentapus-like and in indicates a short metatarsal IV resulting in the eleva-' need of close examination in any comprehensive tion of phalange IV1 above the substrate, and a foot­ survey of theropod tracks from this epoch are the print that is sub-symmetrical, and so atypical for type material Kayentapus hop;; from the Lower Ju­ theropods (Fig. 8). rassic of Arizona (WELLES, 1971), unamed tracks il­ lustrated by LOCKLEY et a/. (1998a) from the Navajo UPPER JURASSIC Formation of southeastern, Utah, tracks from south­ Recent studies of Upper Jurassic tracks in North ern Sweden assigned to Grallator (Eubrontes) America, Europe and Asia have resulted in the for­ soltykovenensis by GIERLINSKI & AHLBERG (1994), mal revision of the ichnogenus Mega/osauripus footprints assigned to Ta/montopus tersi from the (LOCKLEY, MEYER & SANTOS , 1996, 1998). Detailed west coast of France (LAPPARENT & MONTENAT, studies, based on large samples, reveal that this ich­ 1967), Grallatorlescurifrom the region ofthe Massif nogenus has a number of unique morphological Central (DEMATHIEU, 1990) and Schizograllator characteristics which distinguish it in size, track mor­ xiaohebaensis from Yunnan Province China (ZHEN phology and trackway pattern (Fig . 8) from other et a/., 1989). This is to say nothing of various ichno­ known theropod tracks (LOCKLEY, MEYER & SANTOS , taxa from the Lower Jurassic of southern Africa: for 1996, 1998; LOCKLEY, HUNT & LUCAS, 1996; LOCK­ example Neotrisauropus deambulator and Kainotri­ LEY et a/. , 1996). For example the tracks, which sauropus moshoeshoei (ELLENBERGER, 1972, range in length from about 35-75 cm include the larg­ 1974), which may correspond to Kayentapus and est theropod tracks known from the Jurassic (Fig . 8). Eubrontes respectively (sensu WEEMS , 1992). In addition the trackway of Mega/osauripus is re­ In short, as many as five or six ichnospecies ac­ markablywide and irregular, in a large portion of the commodated in four or five ichnogenera (Fig. 7) known sample LOCKLEY, MEYER & SANTOS (1996, could be considered closely allied on morphological 1998), LOCKLEY, HUNT & LUCAS (1996) and LOCK­ grounds. If this conclusion is confirmed, by further LEYet a/. (1996). Some large Mega/osauripus tracks study it would be desirable to revise the ichnotaxon­ lack discrete digital pad impressions, but this ap­ omy of all this material. Such work should only be un­ pears to be a function of preservation in platform car­ dertaken however with reference to the classic bonates, and is not a primary morphological feature, Connecticut Valley collections. Forexample WEEMS since others in the same sample reveal discrete pad (1987, 1992) has already proposed that Kayentapus impressions (see Fig. 8 for comparison between two is an appropriate name for certain material previ­ tracks at opposite ends of the size spectrum). ously assigned to the ichnogenus Apatichnus I n North America and Asia Mega/osauripus co­ (HITCHCOCK, 1858). Without thorough study of the occurs with a smaller theropod track that is distin­ type material the use of the name Kayentapus to re­ guished by its elongate shape and lack of discrete label material from the Hitchcock collection can not digital pad impressions. This ichnogenus has been be supported or rejected. Similarly the fact that Ta/­ labeled Therangospodus (Fig. 8) based on the origi­ montopus predates Kayentapus, should not be seen nal unpublished work of MORATALLA (1993), and as justification for adopting the former name to re­ subsequent observations (LOCKLEY, MEYER & SAN­ place Kayentapus since the type material of Talmon­ TOS, 1996, 1998; LOCKLEY, HUNT & LUCAS, 1996 topus is not abundant in relation to large samples of and LOCKLEY et a/. , 1996).

290 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

~ - .. I E A (.) B ~)C o N

, v ,,) D E

G H

Fig. 7 - Kayentapus-like tracks from the Lower Jurassic, have been assigned a number of ichnotaxonomic labels, some of which are probably synonymous. A - Kayentapus hopiifrom the Lower Jurassic of Arizona (after WELLES, 1971). B, C - Unnamed tracks from the Navajo Formation, Lower Jurassicof Utah (after LOCKLEYet al., in press). D, E - Graflator (Eubrontes) soltykovenensis (from LApPARAENT & MONTE NAT, 1967). G, H - Graflator leseuri (after DEMATHIEU, 1990). 1- Sehizograflator xiaohebaensis (after ZHEN et al., 1989).

291 M. LOCKLEY

LOWER CRETACEOUS to have pad impressions (Fig. 8). They were des­ cribed by SHULER (1935) as Eubrontes (7) g/enro­ There are few examples of detailed studies of sensis, but as pointed out by LOCKLEY, MEYER & large samples of Lower Cretaceous theropod tracks SANTOS (1998) this ichnogenus label is invalid, as from North America. The earliest study, by STERN­ were attempts to assign it tothe ichnogenus Mega/o­ BERG (1932) pertaining to Aptian age tracks from the sauropus (HAUBOLD, 1971). Thus the ichnospecies peace River region of British Columbia resulted in g/enrosensis, has orphan status with regard to its the naming of ichnogenus /renesauripus containing ichnogeneric category. There is no shortage of ma­ two ichnospecies and the monospecific ichnogenus terial in the Texas samples (FARLOW, 1987, PITT­ Co/umbosauripus. (GYPsichnites, although attrib­ MAN, 1989,1992), presumably suggesting the possi­ uted to an ornithopod, may be of theropodan origin bility for a thorough ichnotaxonomic revision and also since Sternberg considered it comparable to comparison of this material with other theropod track Eubrontes). Sternberg's descriptions are brief and assemblages. lack discussions that provide detailed comparisons between these and other tracks. It appears however UPPER CRETACEOUS that these tracks all lack well-preserved digital pad impressions. It is therefore possible that this lack of Two distinctive theropod track morphotypes from discrete pad impressions, as seen in Therangospo­ the Upper Cretaceous of western North America dus, and the Texas tracks described below, is a pri­ have recently been described, and others are mary morphological feature (LOCKLEY, MEYER & known. The first Tyrannosauripus from the Maas­ MORATALLA,1998). trichtian part of the Raton Formation of New Mexico (LOCKLEY & HUNT, 1994) is highly distinctive. It is the Recently the Lower Cretaceous (Berriasian) largest theropod track known, and reveals a marked theropod track Buekeburgichnus maxim us (KUHN, disparity in size between the large well-padded, 1963) has been restudied (LOCKLEY, 1999). It is functional walking digits (II-IV) and the slender hal­ clear that this is a distinctive morphotype with a lux (Fig. 8). Individual phalangeal pad impressions thick, well-padded digit II and a more slender slightly in the predicted formula 2-3-4 are not observed, pre­ segmented digit IV (Fig. 8). This morphology seems sumably owing to the substantial padding of foot to indicate a morphotype different from that found bones that one would expect in a animal weighing as latter in the Cretaceous (e.g., ?Aptian-Albian of much as five to seven tons. However, digit IV is sig­ Spain (MORATALLA & SANZ, 1997) where Buekebur­ nificantly more slender than digit II and in this regard gichnus has also been reported. As noted by the tracks resembles Bueckeburgichnus, though MORATALLA (1993) and WRIGHT et at. (1998) there the central digit is less strongly tapered. are discernable differences between the older theropod ichnites found near the Jurassic­ The second named track Saurexallopus, from Cretaceous boundary, and those higher in the the Harebell Formation of Wyoming (HARRIS et a/., (?Aptian-Albian) succession. Following the restudy 1996; HARRIS, 1997) is also known from the Lance of Buekeburgichnus maximus I consider it unlikely Formation (LOCKLEY et a/., in prep.; Fig. 8 herein). It thatthe large, thick toed (i.e., well-padded) theropod is a highly distinctive track, owing to its very slender tracks from the upper part of the sequence belong to digits and large hallux impression. It also shows a this ichnospecies. They more closely resemble pronounced asymmetry in the hypexes between di­ tracks from the Lower Cretaceous (Albian) of Texas gits II and III and III and IV with the latter being much discussed below. more posterior. Such a configuration indicates that digit II is attached tothe metatarsus in an anterior po­ Studies by Richard McCrea of Albian age thero­ sition to provide space for the attachment of the hal­ pod tracks in western Alberta provide improved de­ lux. This morphology can be contrasted with another scriptions of /renesauripus (Fig. 8) and other Upper Cretaceous theropod track recently discov­ theropod tracks originally described by STERNBERG ered in the Hell Creek Formation (Fig. 4) in which the (1932). In contrast to forms like Eubrontes and indentation between digits II and III is more pro­ Mega/osauripus from the Jurassic, which show nounced and posteriorly located than the indenta­ more or less parallel sided digits and only slightly di­ tion between digits III and IV. Some examples of this vergent toes, /renesauripus, reveals strongly taper­ morphotype reveal a small hallux. Recent discover­ ing digits and widely divergenttoes (100° compared ies of this large slender-toed track from the Hell with only 50-60°). Such differences appear to be of Creek Formation of South Dakota (TRIEBOLD et a/., clear taxonomic significance. 1999) clearly reveal its distinctive morphology. The Large theropod tracks, that lack discrete digital tracks show slender tapering digits, but lack well­ pad impressions (Fig. 8) are well-known from the Al­ defined pad impressions. The hypex between digits bian of Texas, and are somewhat youngerthatthose II and III is much more recessed (posterior) than the from Canada, such as /renesauripus, which appear hypex between digits III and IV, and is therefore the

292 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

Maastrichtian Campanian Santonian Coniacian TV Turonian Cenomanian Albian Aptian Barremian Hauterivian Valanginian Berriasian Tithonian Kimmeridgian Oxfordian Callovian A M Bathonian \\Uc1 Bajocian \) 1---'----1 C Aalenian Toarcian 50 em Pliensbachian Sinemurian Hettan ian .~A .~ r---'-"-----t- Wf ~ Rhaetian G G ~~------I Norian Carnian ladinian Anisian 0-20 20-40 40-60 60-90 em

Fig. 8 - The stratigraphic distribution of distinctive theropod tracks from western North America, with unanmed trackes stippled. G - Grallator. K - Kayentapus. E - Eubrontes. C - Carmelopodus. T - Therangospodus. M - Megalosauripus. B - Bueckeburgichnus (from Germany). 1- Irenesauripus. X - ichnospecies (glenrosensis) of uncertain ichnogeneric affin­ ity. S - Saurexallopus. TY - Tyrannosauripus. XX - Unnamed track from the Maastrichtian of South Dakota. See textforde­ tails.

293 M LOCKLEY

reverse of the configuration seen in Saurexallopus. individual pad impressions, and the degree of asym­ Such fundamental morphological features again metry in hypexes on either side of digit III. If tracks suggest a distinctive ichnotaxon. The track illus­ are not subjected to the whole range of possible trated in Figure 8 is one of a series of 16 in a track­ measures and analyses listed above, it is valid to way, that is thought to best represent the true mor­ ask, wha.t parameters are most important in the in­ phology of the foot. This material is still under study terests of standardizing or normalizing measures of (TRIEBOLD 1999; LOCKLEY & WRIGHT, 1999). variability. Here we return again to the problem of selectivity OTHER CONSIDERATIONS: PRESERVATION in our choice of methodology. As already argued, it is Throughout this discussion I have assumed that neither practical or scientifically necessary to meas­ the practicing ichnologist will follow sensible guide­ ure every possible parameter, or employ every me­ lines and only analyse tracks, for ichnotaxonomic thod used in previous studies. A case can be made purposes if they are reasonably well-preserved that it is necessary to select those parameters mQ§t (BAIRD, 1957; SARJEANT, 1989). In practice, prob­ suitable to the task. In other words certain tracks lems arise when ichnologists study large track sam­ show obvious and characteristic features. The ich­ ples that include poorly preserved tracks. It may nologist should attempt to describe these features then become a difficult and subjective process to and determine whether they are unique ordistinctive separate well-preserved footprints from those that in comparison to other footprints. From a philosophi­ are poorly-preserved. In other words to screen out cal viewpoint, this means treating the track from a "extra-morphologic" variation. Ultimately there are phenomenological perspective (i.e., the distinctive no foolproof criteria available for making such dis­ characteristics of the track influence and in part de­ tinctions. In a given sample the best preservation fine the characteristics that should be measured). may be inferior to that found in the least well-pre­ This means that the observer pays attention to the served tracks in another sample. Conversely what fact that he or she has noted obvious, and real , char­ may appear to be poor preservation, for example the acteristics in their object of study, prior to measure­ lack of claw impressions orthe lack of discrete digital ment and analysis of such characteristics (cf. pad impressions, may be a primary morphological MEDEWAR, 1996). This is in contrast to the ostensi­ feature in some cases. In such cases it would be bly analytical approach that seeks to "objectively" wrong to disregard tracks on the assumption that measure as many parameters as possible, deliber­ they were poorly preserved. The only way to begin to ately overlooking any "subjective" observations of address such problems is to try and familiarize one­ distinctive or special features, hoping instead to self with, and describe, large samples, and establish quantify purportedly objective mor-phological char­ criteria for recognizing the optimum state of preser­ acteristics according to mathematical or statistical vation. Within large samples that show atleast some models. This latter approach assumes that the ob­ well-preserved tracks it is then desirable to deter­ server of obvious characteristics may be mistaken in mine whether certain characteristic features are their perceptions. More often than not however, the seen repeatedly. This can be done by superimpos­ analysis will confirm that obvious "qualitative" char­ ing multiple tracings (BAIRD, 1957). acteristics can very easily be defined in quantitative terms, thus disproving the hypothesis that the ob­ SUMMARY PERSPECTIVES ON VARIABILITY server was blind or deluded in the first place. IN THEROPOD TRACK MORPHOLOGY It is also evident that morphological, and hence As indicated above, theropod track morphology taxonomic variability in fossil footprints is greater is quite variable (Fig. 8). Regardless of the extent to through time than in space. In other words, ichno­ which morphological differences have been descri­ taxa tend to have restricted stratigraphic ranges, but bed or quantified using the many criteria listed extensive geographical distributions (cf., HAUBOLD above, no one could confuse Grallator with Tyran­ & KATZUNG, 1978; LOCKLEY, 1998; LOCKLEY, MEYER nosauripus, or Saurexallopus with Mega/osauripus & SANTOS, 1996), as one would expect from a gen­ on simple inspection. Only a limited number of mea­ eral knowledge of the fossil record and evolution. Al­ surements would be required to provide the type of though this rule is by no means infallible, it does differential descriptions already published with re­ suggest that ichnologists should not expect to find spectto these ichnogenera. It is clear from simple in­ extensive, taxonomically-significant morphological spection ofthe range of variability seen in the tracks variability within restricted stratigraphic intervals. In illustrated in Figure 8, that such variability can effec­ this regard there is apparently an artificially high tively be described in terms of parameters such as peak ofichnotaxonomicdiversity in the Lower Juras­ shape ("elongateness"), relative width ("slender­ sic owing to the early work of American ichnologists ness") and tapering of the digits, presence or ab­ (H ITCHCOCK, 1858; LULL, 1953) and subsequent sence of a hallux and creases that separate work on European and South African ichnofaunas,

294 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

(LAPPARENT & MONTENAT, 1967; ELLENBERGER, to make connections between track and potential 1972,1974; DEMATHIEU, 1990, GIERNLINSKI , 1991). trackmaker should the researcher so desire. This Since detailed justification for the introduction of endeavor, however, carries with it the risk of attach­ new names is limited or lacking in many of these ing an inappropriate label to a particular track type. studies, it can be argued that this peak is artificial But this risk; iA and of itself, does not mean that dis­ and does not reflect increased taxonomic diversity, tinctive ichnotaxa should not have appropriate even though theropods may have been undergoing names, only that we face a dilemma in choosing a a significant radiation at this time. suitable name.

DO WE NEED TO KNOW THE TRACK MAKERS? TRACKS IN THEIR OWN RIGHT: THE PHENOMENOLOGICAL APPROACH. According to SMITH & FARLOW (1996: 46) "( ... ) paleoecology and biostratigraphy using footprint ta­ The perception that tracks, particularly those of xa require confident assignment of trackmakers to theropods, provide limited information, is an unwar- _ the tracks." This is not valid at all, and reflects mud­ ranted scientific assumption, and, if it leads to lack of died thinking about biostratigraphy. One can do bi­ study, then tends to become a selffulfilling prophecy. ostratigraphy without knowing the affinity of the in­ The considerable morphological variability, remark­ dex fossils, as was done with conodonts for many able abundance and distinctive distribution patterns years. While it is true that one needs to know the af­ recorded for theropod tracks, in recent years sug­ finity of a trackmaker in order to more fully interpret gest that, like all tracks, they are worth studying in the paleoecology, one can go a long way towards their own right as a valuable source of information. furthering knowledge of ichnofaunas and ichnofa­ This opportunity is missed if there is a perception cies, simply by describing and defining ichnotaxa, that tracks are only useful once the trackmaker has and ichnotaxonomic assemblages. This is, in part, been identified. With this in mind the theropod track the ichnofacies approach, which does not require a examples cited herein from the western United knowledge ofthe tracemakerto proceed. For exam­ States (Fig. 8) can be seen, in all cases to provide ple we have recognized the Carmelopodus assem­ new morphological information about theropod feet blage in the Middle Jurassic of Utah and England (and locomotor behavior), regardless of whether we (LOCKLEY et al., 1998b), and have defined the Mega­ are confident in interpreting the affinity of the track­ losauripus- Therangospodus association in the lo­ makers. wer part of the Upper Jurassic of North America, As indicated in the section on the track record, Asia and Europe (Fig. 8). While all three ichnoge­ such new morphological information justifies the as­ nera are interpreted as theropodan in affinity, only signment of new taxonomic labels. It should be the Megalosauripus is attributed to a particular taxo­ primary business of any morphological study to rec­ nomic group atthefamiliallevel, and this without any ognize, and intelligently describe significant mor­ certainty. A change in the interpretation of the affinity phological features of a species or taxon, regardless of one or more of these ichnogenera would not alter of preconceived ideas about variability within taxo­ the importance of the assemblage as a reflection of nomic groupings. Such principles should, and do, the distribution of distinctive vertebrate track mor­ apply equally to body and trace fossils. Once a spe­ phologies in space and time. Similarly the recogni­ cies has been described, on the basis of unique, tion of tracks such as Tyrannosauripus and Sau­ ratherthan shared or common characteristics, it can rexallopus in the Cretaceous are useful for compari­ be compared more effectively with other species. son of track assemblages within western North Such principles again stress the importance oftreat­ America, even though in the former case we claim to ing tracks as separate entities from potential track­ have identified the trackmaker, and in the latter case makers and analysing their morphology in their own we do not. right. Given that tracks represent only a part of the A clear distinction should always be made be­ trackmaker's anatomy, there is even more reason to tween the morphological - i.e., taxonomic- descrip­ thoroughly examining all aspects of track morphol­ tion of tracks and the identification oftrackmakers. If ogy. In the examples given from the western U.S.A. the two activities are integrated it must be done as a (Fig. 8) it is argued that the tracks have been exam­ two step process The track should first be described ined and described from this essentially phenome­ on the basis of its morphological characteristics, nological perspective (i.e., with the focus on unique with the decision as to whether it is unique and wor­ and distinct features that facilitate differentiation). thy of a taxonomic label based only on morphologi­ This approach seeks to reveal diversity in unity, cal uniqueness. Once this decision has been made rather than unity in diversity, (cf. BORTOFT, 1996). the choice of a name can be based on inferred affin­ This philosophical concept can be explained heuris­ ity to an osteological species, though other criteria tically as follows. If we take a holistic overview of all may be used. At this stage every effort can be made known morphologies in a particular field (theropod

295 M. LOCKLEY

tracks), we will observe a range of distinctive mor­ tyl or theropod ichnogenera found in a given for­ phologies (the entire field) centered on, or diverging mation or palichnostratigraphic zone. from, more generalized and less distinctive types (Fig. 9). These latter types are often the smaller Ich­ Such an overview soon leads again to the observati­ nospecies. Large tracks, althe periphery of this field on that Lower Jurassic theropod taxonomy, arising will stand out as quite distinctive, because disparity from assemblages in the eastern United States is or variation from the mean appears to increase with confused, for historical reasons (OLSEN et al., 1998). size, being in part a function of the way morpho­ Thus these track assemblages from this region (and space is occupied (FOOTE, 1991). The larger forms epoch) can not be fully understood in their geo­ are, in effect distinguished from the remainderofthe graphic, stratigraphic and taxonomic context, sample by a process of eli mination -essentially start­ without considerable detailed historical research . ing althe periphery and working. Put another way, It By contrast track assemblages from other epochs is legitimate to look for natural differences (dispari­ and regions (e.g., the Middle Jurassic to Late Creta­ ties) among that part of the sample (ichnospecies) ceous of the western United States), are of known where they are obvious, before trying to differentiate geographical and stratigraphic origin, and straight­ morphologies in samples where disparity is subtle forward taxonomic status, uncomplicated by a com­ and hard to recognize. plex history. Moreover they have been studied and described with reference to large samples. The contrasting approach seeks to emphasize and define common features by imposing unity on Thus it can be concluded, from direct observa­ diversity, sometimes masking or de-emphasizing tion that there is a diversification of theropod track distinctive features. Such perspectives highlight the mor'phology through time. The small generalized merits of seeking disparate or distinctive features and ubiquitous Late Triassic and Early JurassIc rather than common features. Entities such as pale­ ichnogenus Gral/ator, already accompanied by Eu­ ontological species, or ichnospecies are recognized brontes and Kayentapus in the early Jurassic, does by their distinctiveness not by their similarity. Such not persist abundantly throughout the Mesozoic, simple recognition of distinctive morphologies, though sporadic post-Early Jurassic examples may which has formed the basis of traditional clasSifica­ tions, can and should be backed up by suitable mea­ surements and descriptions. Forexample in the mo­ del shown (Fig. 9) peripheral end members of the LONG ~ morphological plexus are more easily differentiated from the mean than those clustered close together. Therefore the number of parameters needed to dif­ ferentiate tracks types is to some degree inversely GRACILE proportional to the distance between points of com­ ROBUST parison. Those tracks that are separated from oth­ ers by natural breaks in the morphological conti­ 6 IE---- nuum will naturally stand out on the basis of a few distinctive characteristics, whereas those clustered close to the center of the field will require more so­ phisticated and subtle means of discrimination. Other concrete realities that are informative in their own right pertain to the observed structure of the track record (i.e., the distribution of morpholo­ gies in space, time and in relation to facies). Such features of the track record can also be described and understood regardless of the known affinity of ) the trackmakers. A primary characteristic is that increased deviation tracks are very abundant, and often provide large from mean samples consisting of only a few distinctive taxa from any given stratigraphic level. Arising from these Fig. 9 - Schematic model of-multidimensional morpho­ observations we begin to see that most stratigraphic logical variation in theropod tracks, showing how end intervals of stage magnitude, or age duration, con­ members are easily differentiated from the mean. Model tain distinctive ichnofaunas, but rarely is the diver­ assumes both a constant (linear) morphological gradient and natural breaks in the morphological continuum. The sity of ichnofaunas exceptionally high. In other two axes, transverse to elongate (north to south) and ro­ words rarely are there more than two or three tridac- bust to gracile digit widths (east to west) are arbitrarily cho­ sen as typical variables , and are entirely schematic.

296 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

exist. The appearance of Carme/opodus (with re­ is the size and apparent uniqueness of the sample duced metatarsal IV), Therangospodus with coa­ and how does it fit into the local regional or global lesced pads and Megalosauripus with tendencies stratigraphic context. Are there many existing ichno­ towards gigantism and a wide, irregular trackway in­ taxa know from this time? Are they variable, homo­ dicate a variety of derived characteristics that are geneous, well-known, problematic or controversial? clearly reflected in the theropod track record, but not Will discrimination be easy on the basis of simple with the same degree of explicitness in the bone rec­ and few measures, or is a detailed study antici­ ord. Similar distinctive characteristics show up in the pated? In short, a taxonomic preamble is desirable Cretaceous track record, notably with Saurexal­ to introduce and justify the type of taxonomic study lopus and Tyrannosaripus, but also with Bueckebur­ appropriate to the material. Then the methods used gichnus and Irenesauripus and certain unnamed can be more convincingly justified. All descriptive, forms. It remains to be seen to what extent these qualitative, semi-quantitative and quantitative meth­ tracks stand out from others not yet named or dis­ ods are potentially valid, but some may be more ap­ covered (i.e., will we find additional ichnotaxa in the propriately employed than others in certain Cretaceous that are indisinguishable from known circumstances. Simply to measure everything with Jurassic ichnotaxa, and will such ichnotaxa form a methods described as "sophisticated" or "powerful" significant component of the fauna? The answer to may be as much a test of a complex methodology as this latter question is probably no, in most cases). it is a readily understandable method of discriminat­ ing and differentiating primary morphological fea­ The present trend , at least for western North tures, which may be recognizable and measurable America, is for new discoveries to point to a substan­ by a variety of means. This is not to suggest that it is tial evolutionary diversity oftheropod track morphol­ inappropriate to apply complex or sophisticated ogy through time, reflecting in some cases, the rise measures, only that practicioners of such methods of well-known groups (e.g., megalosaurids and ty­ should explain the morphological and ta xonomic rannosaurids), while in other cases demonstrating significance of their results. the abundance of theropod groups not yet confi­ dently matched with their tracks. As the track record Given the evidence of over-splitting in studies of becomes better known and comprehensively de­ Early Jurassic tracks, it is hard to justify the introduc­ scribed, we can predict that more correlations be­ tion of new names, without revision of existing ichno­ tween tracks and skeletal remains will become taxonomic problems, unless of course, entirely new possible. This process will be facilitated by a study of and highly distinctive morphologies are discovered. the stratigraphic range of ichnotaxa and potential By contrast studies of tracks from other epochs sug­ trackmakers, not just by morphological analysis. gest that there may be somewhat more need and justification for introduction of new taxonomic na­ DISCUSSION, CONCLUStONS AND mes simply because many apparently new, or previ­ RECOMMENDATIONS ously unstudied track types lack formal descriptions. Here again it is proposed that we tailor or modify our In any taxonomic study of fossil footprints there ichnotaxonomic perspective to suit the reality of the are minimum standards of procedure that should be problem we are dealing with. followed (SARGEANT, 1989). These standards do not require detailed, multiple quantitative analyses Growing palichnostratigraphic evidence now of the type listed above, though the use of suitable emphasises the need to see track assemblages in quantitative measures is recommended . The crucial their proper geologic and evolutionary contexts. issue is therefore subjective with regard to what fea­ Tracks should first be compared with footprints from tures are "suitable ." Here we can return to first princi­ rocks of the same age, before extending the tempo­ ples and remind ourselves of the recommendation ral range of investigations. Similarly, search for po­ of BAIRD (1957) that we base serious and formal tential trackmakers should be conducted primarily in taxonomic work on samples that contain adequate coeval deposits. This significantly reduces the num­ supplies of well-preserved material. The recent ren­ ber of potential permutations, and minimizes the aissance in vertebrate ichnology has benefited from temptation to make unparsimonious diachronistic discovery of much new material, especially in the correlations. The contrasting approach of compar­ western United States, that fits these criteria, and ing tracks from different periods, epochs and ages is has allowed for the description of new ichnotaxa fraught with potential and real problems. This is es­ based mainly on substantial samples of known geo­ pecially true in the case of theropod tracks. For ex­ graphic and stratigraphic origin. ample, until recently, other than Grallator and Eu­ brontes, from the Lower Jurassic of New England, Given that all samples have a context it is consid­ there were few well-documented, named tracks with ered important to preface any taxonomic discussion which other theropod tracks (e.g., the famous Creta­ or description with a statement of the problem. For ceous tracks of Texas), could be compared . Thus, example: why does the sample warrant study. What

297 M. LOCKLEY

when cursory studies were undertaken, they tended pie, though only two named forms are from North to be given the same or similar ichnotaxonomic la­ America (LOCKLEY & RAINFORTH, in press). Thus bels (e.g., ?Eubrontes), thereby "creating" the per­ there is no evidence that morphological variation of ception that all theropod tracks tend to be similar up theropod tracks is significantly lower than in ornitho­ and down the stratigraphic column, when in reality pods, the only other group of bipedal dinosaurs that no one has looked closely at the problem. have a comparable stratigraphic range and general morphology. On the contrary, based on the number Evidence presented herein, refutes or at least of named ichnotaxa theropod track variation may be undermines this generalized assumption, in many greater than it is among the ornithischians. cases, and highlights the real perceptual problems that arise from being too eager to put everything into ACKNOWLEDGMENTS the same category, to satisfy the preconception that this is an objective or scientific approach. This re­ The research on which this paper is based was duction of diversity into unity then requires the appli­ conducted over many years, with the support of the cation of highly discriminating analyses of multiple Offices of the Dinosaur Trackers Research Group at parameters to disentangle samples that it may not the University of Colorado at Denver. Primary fund­ have been necessary to mix in the first place. Such ing, between 1987 and 1997, was provided by the an approach is not necessarily wrong so much as National Science Foundation. I thank various col­ unnecessary, if simpler observational methods and leagues, who are listed as my co authors in the bibli­ geographical and stratigraphic sample differentia­ ography, for miscellaneous help with field work, tion are appropriately acknowledged in the first pla­ discussion and documentation. I also thank Joaquin ce. Similarly, in the game of seeking potential track­ Moratalla and Steve Gatesy for their reviews of the makers from all over the stratigraphic column, not manuscript. only is the number of potential trackmakers increa­ sed unnecessarily, but such potential trackmakers, REFERENCES for the most part, become theoretical, rather than BAIRD, D. (1957) - Triassic reptile footprint faunules from Milford, real possibilities, because many of them did not exist New Jersey. Bull. Museum Compo Zool., 117: 449-520. at the time that the tracks were made. Such an ap­ BORTOFT, H. (1996) - The wholeness of nature: Goethe's way of proach is confusing and has significant conceptual science. Lindisfarne Press, New York, 407 pp . problems for ichnological reality. It is in a very real CURRIE, P.J . & SARJEANT, W.A.S. (1979) - Lower Cretaceous di­ sense a tendency towards blind faith in the analytical nosaur footprints from the Peace River Canyon, Britsih Co­ lumbia, Canada. Pafeogeog., Palaeoclimato'" Palaeoeco/. , methods of science, atthe expense of careful obser­ 28: 103-115. vation of the objects of study. CURRIE, P.J .; NADON, G. ; LOCKLEY, M.G. (1991) - Dinosaur Fo­ otprints with Skin Impressions from the Cretaceous of Alberta A final point worth consideration in our discus­ and Colorado. Can. J. Earth Sci., 28: 102-115. sion of morphological diversity in the theropod track OEMATHIEU , G. (1990) - Problems in the discrimination of tridactyl record, is an evaluation of how it compares with the dinosaur footprints , exemplified by the Hettangian trackways , record of other tridactyl bipeds such as ornithopods the Causses, France . lchnos, 1: 97-110. and birds. It is beyond the scope of this paper to go OEMATHIEU, G. (1993)- Empreintes de pas de dinosaures dans Ie into this too deeply, or to discuss the extent to which Causses (France). Zubia, 5: 229-252. problems may arise in trying to differentiate tracks OEMATHIEU , G. & SCIAU, J. (1995) - L'ichnofaune Hettangienne O'archosauriens de Sauciieres, Aveyron, France. BUll. Soc. attributable to these three groups. Based on current Hist. Nat. Autun, 151: 5-46. knowledge, however, the only widely-accepted, ELLENBERGER, P. (1972) - Contribution a la classification des piste widely-distributed named ornithopod tracks are de vertebres du Trias : les types du Stormberg d'Afrique du Anomoepus (probably a senior synonym of Moye­ Sud (1). Palaeovertebrata, Mem. Extraordinaire, 116 pp. nosauripus) and Dinehichnus from the Jurassic ELLENBERGER, P. (1974) - Contribution a la classification des pis­ (HITCHCOCK, 1958; LULL , 1953; LOCKLEY et aI., tes de vertebres du Trias; les types du Stormberg d'Afrique du 1998c) and so called "Iguanodon" tracks, Ambly­ Sud , 2e partie. Palaevertebrata, Mem. Extraordinaire, 170 pp. dactylus, Caririchnium, and unnamed hadrosaur FARLOW, J.D. & LOCKLEY, M.G. (1993) - An Osteometric Appro­ ach to the Identification of the Makers of Early Mesozoic Tri­ tracks from the Cretaceous (CURRIE & SARJEANT, dactyl Dinosaur Footprints. New Mexico Museum Nat. Hist. 1979; CURRIE et al., 1991; LOCKLEY & HUNT, 1995). Sci. Bufl., 3: 123-131 . This total of five formally named and two informally FARLOW, J.D. & CHAPMAN , R (1997)- The Scientificstudyofdino­ named ichnogenera from western North America, is saur footprints, in FARLOW, J.O & BRETT-SURNAM , M.K. (Eds.), The Complete Dinosaur, Indiana Univ. Press , only about half the total for distinctive Jurassic and Bloomington , pp. 519-553. Cretaceous theropod tracks described herein from FOOTE, M. (1991) - Morphological patterns of diversification: essentially the same stratigraphic sections. Bird examples from trilobites. Paleontology, 34: 461-485. tracks are known only from the Cretaceous, com­ GATESY, S.M.; MIDDLETON, K.M.; JENKINS, F.A.JR. & SHUBIN, prising at least eight named ichnogenera, and a few N.H. (1999) -Three- dimensional preservation of foot move­ other distinctive unnamed forms in the global sam- ments in Triassic theropod dinosaurs. Nature, 399: 141-144.

298 PHILOSOPHICAL PERSPECTIVES ON THEROPOD TRACK MORPHOLOGY

GIERLlNSKf, G. (1991) - New dinosaur ichnotaxa from the early Ju­ LOCKLEY, M.G.: MEYER, C.A. & MORATALLA, J.J. (1998) - Theran­ rassic of the Holy Cross Mountains, Poland. Pa/eogeog., Pa­ gospodus: trackway evidence for the widespread distribution leoclimato/., Paleoeco/., 85: 137-148. of a Late Jurassic theropod with well-padded feet. Gaia (this volume). GIERLINSKI, G. & AHLBERG, A. (1994) - Late Triassic and Early Ju­ rasic dinosaur footprints in the Hoganas Formation of LOCKLEY, M.G.; MEYER, C.A. & SANTOS, V.F. (1998) - Megafosau­ southern Sweden./chnos, 3: 99-105. ripl.l.s, and~ the problematic concept of Megalosaur footprints. Gaia (this volume). GOODWIN, B. (1994) - How the leopard changed its spots: the evo­ lution of complexity. Touchstone Books. Simon & Schuster, LOCKLEY, M.G.: HUNT, A.P.; MEYER, C.M.; RAIN FORTH, E. & New York, 252 pp. SCHULTZ, R. (1998a) - A survey offossil footprints at Glen ca­ non National Recreation Area (western USA): a case study in HARRIS, J.D.; JOHNSON. K.R. ; HICKS, J. & TAUXE, L. (1996) - Four­ documentation of trace fossil resources. Ichnos, 5: 177-211 toed theropod footprints and a paleomagnetic age from the Whetstone Falls Member of the Harebell Formation (upper LOCKLEY, M.G.; HUNT, A.P.; PAQUETIE, M.; BILBEY, S-A. & HAM­ Cretaceous: Maastrictian). northwestern Wyoming. Gret. BLIN, A. (1 998b) - Dinosaurtracks from the Carmel Formation, Res., 17: 381-401. northeastern Utah: implications for Middle Jurassic paleoeco­ logy. fchnos, 5: 255-267. HARRIS, J.D. (1997) - Four-toed theropod footprints and a paleo­ magnetic age from the Whetstone Falls Member of the Hare­ LOCKLEY, M.G.; SANTOS, V.F.; MEYER, C. & HUNT, A.P. (1998c)­ bell Formation (upper Cretaceous: Maastrictian). northwes­ A new dinosaurtracksite in the Morrison Formation, Boundary tern Wyoming: a correction. Gret. Res., 18: 139. Butte, Southeastern Utah, in CARPENTER, K.; CHURE, D. & KIRKLAND, J. (Eds.), The Upper Jurassic Morrison Formation: HAUBOLD, H. (1971) - fchnia Amphiborum et Reptiliorum Fossi­ an interdisciplinary study. Modern Geo/., 23: 317-330. fum. Handbuch der Palaeoherpetologie Teil 18. Gustav Fischer Verlag, New York, 124 pp. LOCKLEY, M.G. (1999a) - The Eternal Trail: a tracker looks at evo­ lution. Perseus Books, Reading, Massachusetts, 334 pp. HAUBOLD, H. & KATZUNG, G . (1978) - Paleoecology and paleoen­ vironments of tetrapod footprints from the Rotliegend (Lower LOCKLEY, M.G. (1999b) - Einblicke in die Gestaltebiofogie derDi­ Permian) of Central Europe. Pa/eogeog., Paleoclimatol., Pa­ nosaurier anhand ihrer Fahrtenspuren. Tycho de Brahe­ leoeco!., 23: 307-323. Jahrbuch fur Gotheanismus, Nierfern~Oschelbronn, pp. 134- 166. HAUBOLD, H. (1984) - Saurierfahrten. (2nd Edition) Wittenberg Lutherstadt, Germany, Die Neue Brehm-Bucherei, 232 pp. LOCKLEY, M.G. & MEYER, C.A. (1999) - Dinosaur tracks and other Fossil footprints of Europe. Columbia Univ. Press, New York, HITCHCOCK, E. (1858)- Ichnology of New England. A report on the 323 pp. Sandstone of the Connecticut Valley, especially its Fossil Fo­ otmarks. W. White, Boston (reprinted 1974 by Arno Press, LOCKLEY, M.G. & WRIGHT, J . (1999) - Upper Cretaceous (Maastri­ New York), 232 pp. chtian) tracksites from the Western Interior, USA: a sythesis. J. Vertebr. Paleontol., 19: 59A. IRBY, GV. (1995) - Posterolateral markings on dinosaur tracks, cameron dinosaur tracksite, Lower Jurassic Moenabe Forma­ LOCKLEY, M.G.; NADON , G.; WRIGHT, J. & CURRIE, P. (In prep.)­ tion, Northeastern Arizona. J. Paleontol., 69: 779-784. Dinosaur and Bird footprints from the Lance Formation, eas­ tern Wyoming. KUHN, O. (1963) - Ichnia Tetrapodium, in WESTPHAl. F. (Ed.), Fossilium Cata/ogus 1: pars 101, Junk s'Gravenhage, 176 pp. LOCKLEY, M.G. & RAINFORTH, E. (in press) - The tracks record of Mesozoic Birds and Pterosaurs: an ichnological and paleoe­ LAPPARENT, A.F. DE & MONTENAT, G. (1967) - Les empreintes des cological perspective, in CHIAPPE, L & WITMER, L.M. (Eds.), pas de retiles de I'infralias du Veillon (Vendee). Soc. Geo/. Mesozoic Birds Above the heads of dinosaurs, Univ. Califor­ France, Mem. n. s., 107: 1-41. nia Press. LEONARDI, G. (1987) - Glossary and manual of Tetrapod Footprint LULL, R.S. (1953) - Triassic Ufe of the Connecticut Valley. Con­ Paleoichnology. Dep. Nac. ProdU/;ao Min. Brasil, 75 pp., XX necticut State Geol. Nat. His!. Survey Bull., 81: 1-33. Pis. MEDAWAR, P .B. (1996) - The strange case of the spotted Mice and LOCKLEY, M.G. (1986) - Dinosaur Tracksites: A Field Guide Pu­ other classic essays on science. Oxford Univ. Press, Oxford, blished in Conjunction with the First International Symposium 236 pp. on Dinosaur Tracks and Traces. Univ. Colorado Denver. Geo/. Dept. Mag., Sp. Issue 1: 1-56 pp. MORATALLA, J.J.; SANZ, J.L. & JIMENEZ, S. (1988) - Multivariate analysis on Lower cretaceous dinosaur footprints: discrimina­ LOCKLEY, M.G. (1991) - Tracking Dinosaurs: a New Look at an An­ tion between ornithopods and theropods. Geobios, 21: 395- cient World. Cambridge Univ. Press, Cambridge, 238 pp. 408. LOCKLEY, M.G. & HUNT, A.P. (1995) - Dinosaur Tracks and other MORATALLA, J.J. (1993) - Restos indirectos de dinosaurious del fossil footprints of the western United States. Columbia Univ. registro espanol: paleoichnologica de la Cuence de cameros Press, New York, 338 pp. (Jurasico superior-Cretacico inferior) y Paleoologia del Creta­ LOCKLEY, M.G.; HUNT, A.P. & LUCAS, S.G. (1996) - Vertebrate cico superior. Thesis doctoral, FacuJdad de Ciencias Univer­ track assemblages from the Jurassic Summerville Formation sida Autonoma, Madrid, 729 pp. and correlative deposits, in MORALES, M. (Ed.), Continental MORATALLA, J.J. & SANZ, J .L. (1997) - Cameros Basin Mega­ Jurassic Symposium Volume, Museum of Northern Arizona, tracksite, in CURRIE, P.J & PADIAN, K. (Eds.), Encyclopedia of Flagstah, pp. 249-254. Dinosaurs, Academic Press, San Diego, pp. 87-90. LOCKLEY, M.G.; MEYER, C.A. & SANTOS, V.F. (1996) -Megalosau­ OLSEN, P.E. (1980)- Fossil great lakes of the Newark Supergroup ripus, MegaJosauropus and the concept of Megalosaur fo­ in New jersey, in MaNSPEIZER, W . (Ed.), Field Studies of New otprints, in MORALES, M. (Ed.), Continental Jurassic Jersey Geology and Guide to field trips, New York State Geo­ Symposium Volume, Museum of Northern Arizona, Flagstah, logical Association, 52nd Annual Meeting, Rutgers University, pp.113-118. pp.352-298 LOCKLEY, M.G.; MEYER, C.A.; SCHULTZ-PITIMAN, R & FORNEY, . G. (1996) - Late Jurassic dinosaur tracksites from Central OLSEN, P.E. & BAIRD, D. (1986)- The ichnogenus Atreipus and its significance for Triassic biostratigraphy, in PADIAN K. (Ed.), Asia: a preliminary report on the world's longest trackways, in The beginning of the Age of Dinosaurs, Cambridge Univ. MORALES, M. (Ed.), Continental Jurassic Symposium Volu­ Press, Cambridge, pp. 61-87. me. Museum Northern Arizona, pp. 271-273 OLSEN, P.E.; SMITH, J.B. & MCDONALD, N.G. (1998) - Type mate­ LOCKLEY, M.G. (1998) - The vertebrate track record. Nature, 396: rial of the type species of the classic theropod footprint genera 429-432. Eubrontes, Anchisauripus and Grallator (Early Jurassic,

299 M. LOCKLEY

Hartford and Deerfield basins, Connecticut and Massachu­ setts, USA). J. Vertebr. Paleontol. , 18: 586-601. PITTMAN, J.G. (1989) - Stratigraphy, Lithology, Depositional Envi­ ronment, and Track type of Dinosaur Track-bearing Beds of the Gulf Coastal Plain, in GILLETTE, D.O. & LOCKLEY, M.G. (Eds.). Dinosaur Tracks and Traces, Cambridge Univ. Press, Cambridge, pp. 135-153. PITTMAN, J,G. (1992) - Stratigraphy and Vertebrate Ichnology of the Glen Rose Formation, Western Gulf Basin, USA. PhD Thesis, University of Texas at Austin, 726 pp. RAsSKIN-GUTMAN, D.; MORATALLA, J.J.; SANZ, J. L & CHAPMAN, R.E. (1996) - The problems and potentials in studying tridactyl footprints, inWOLBERG. D. & STUMP, E. (Eds.), Dinofest Inter­ national Symposium (Abstracts). Arizona State University, Philadelphia, p. 91 RASSKIN-GUTMAN, D.; HUNT, G.; CHAPMAN, R.E.; SANZ, J.L & MORATALLA, J.J. (1997) - The shapes of tridactyl dinosaur fo­ otprints: procedures, problems and potentials; in WOLBERG, D.L.; STUMP, E. & ROSENBERG, G.D. (Eds.), Dinofest Interna­ tional Symposium, Proceedings. A publication of the Natural Academy of Sciences, pp. 377-383

SARJEANT, W.A.S. (1989) - ~ Ten paleoichnological command­ ments ~: A standardized procedure for the description of fossil vertebrate footprints in GILLETTE , D.O. & LOCKLEY, M.G. (Eds.), Dinosaur Tracks and Traces, Cambridge Univ. Press, Cambridge, pp. 369-370. SCHAD, W. (1971) - Towards a Biology of Form. Waldorf Press, New York, 309 pp. SHULER, E.W. (1935) - Dinosaurtracks mounted in the bandstand at Glen Rose, Texas. Field and Lab. , 5: 34-36. SMITH, J. & FARLOW, J.O. (1 996) - Were the trackmakers forthe di­ nosaur ichnotaxa GraJJator, Anchisauripus and Eubrontes really theropods?, in TORNEAU, P.M & OLSEN , P .E. (Eds.), As­ pects of Triassic-Jurassic Rift Basin Geoscience, State Geo­ logical and Natural History Survey of Connecticut, Miscellaneous reports, pp. 46-47. STERNBERG, C.M. (1932)- DinosaurTracks from Peace River Bri­ tish Columbia. Ann. Rept. Nat. Museum Can., (for 1930): 59- 85. THULBORN , R.A. (1990) - Dinosaur Tracks. Chapman Hall, Lon­ don, 410 pp. TRIEBOLD, M .; LOCKLEY, M.G.; JANKE, P . & NYSTROM, C. (1999)­ Dinosaur tracks from the Hell Creek Formation, Upper Creta­ ceous, Maastrichtian) South Dakota. J. Vertebr. Paleontol., 19: 81A. WEEMS, R.E. (1987) - A Late Triassic footprint fauna from the Cul­ peper Basin Northern Virginia (U.S.A.). Trans. Am. Phil. Soc., 77: 1-79. WEEMS, RE. (1992) - A re-evaluation of the taxonomy of Newark Supergroup saurischian dinosaur tracks, using extensive sta­ tistical data from a recently exposed tracksite near Culpeper Virginia, in SWEET, P.C. (Ed.), Proc. 26th Forum on the Geo­ logy of Industrial Minerals. Virginia, Division of mineral resour­ ces, Charlottesville, pp. 113-127 WELLES, S.P. (1971) - Dinosaur Footprints from the Kayenta For­ mation of Northern Arizona. Plateau, 44: 27-38. WRIGHT, J.L.; BARREn, P.M .; LOCKLEY, M.G. & COOK, E. (1998)­ A review of early Cretaceous terrestrial vertebrate track-bea­ ring strata of England and Spain, in LUCAS, S.G.; KIRKLAND , J.1. & ESTEP, J.W. (Eds.), Lower and Middle Cretaceous Ter­ restrial Ecosystems. New Mexico Museum Natural Hist. Sci. Bull., 14: 143-153, ZHEN, S.; lI, R.; RAo, C.; MAnER, N. & LOCKLEY, M .G. (1989) - A Review of Dinosaur Footprints in China, in GILLETTE, D.O. & LOCKLEY, M.G. (Eds.), Dinosaurs Past and Present, Cambrid­ ge Univ. Press, Cambridge, pp. 187-198.

300