TAPHONOMY

Formation of the Conglomerates

Fossiliferous, nodule-rich conglomerates form in areas where pedogenic nodules and vertebrate remains are a substantial part of the coarse fraction of sediments available; i.e., they are likely to be prominent in fine-grained systems in regions where calcareous soils tend to form, and where rates of deposition are comparatively slow, allowing good development of soils. Kraus (1997) reported better developed paleosols in Eocene alluvial beds of the Willwood Formation in Wyoming in areas estimated to have had a 0.3 and 0.4 mm/yr rate of rock accumulation than for areas in which the rate of accumulation was estimated between 0.6 and 0.7 mm/yr. Kraus and Bown (1993) estimated the Willwood rates of accumulation. In the late Paleocene, the Big Bend area received much slower deposition and is less fossiliferous than late Paleocene sites farther north like those of Wyoming, some of which received over 25 times more sediment per million years (Sloan 1987; Schiebout 1995). It is possible that the breaking of bones and teeth, fragments of which are common in the screening residues, has been facilitated by the nodule development itself. Porous bone tends to become heavily encrusted and/or broken where nodules are forming (Figure 31). This has resulted in a generally poor paleontological record for areas of slow deposition and nodule-rich soils, in part because of damage to the fossils and in part because of problems involved in getting the nodular material off of the fossils.

Concentration of nodules and bones can result from reworking during channel meandering, overbank fluvial erosion due to flooding, avulsion, or erosion resulting from a drop in base level that causes rivers to entrench, overbank addition to interfluvial areas to slow or stop, and erosion to begin to denude the interfluves. The conglomerates of very limited lateral extent, like Joe's Bonebed conglomerate from the Paleocene deposits of Big Bend or small lenses exposed at the DISC Site at Fort Polk (Figure 22), probably represent material washed into small channels or low areas on the floodplain. The Cretaceous Aguja conglomerates are examples of lag deposits in larger channels. Wells (1983) discussed pedogenic nodule lag deposits that included some fish and mammal material, from the bases of small channels in sand-poor redbeds from the Eocene deposits of northern Pakistan, and that he interpreted as formed by short-lived, briefly active streams, in an environment analogous to parts of modern southern and central Australia.

The most laterally extensive deposit at DISC site in the Louisiana Miocene represents a locally variable blanket of material concentrated from soils by erosion. The thin mudstone layer between upper and lower portions of the main conglomerate may represent an avulsion or crevasse splay event flooding part of the area or may represent local sheet wash. The extreme fragility of some of the fossils recovered from the DISC site precludes long transport, but the well-rounded condition of many bone pieces (particularly those from larger animals), suggests some transport for them, perhaps including several cycles of reworking. A larger scale example of similar processes has been reported from the Lower Jurassic of South Africa, where regional degradation resulting from a base level drop produced a pedogenic nodule concentrate 0.5-1.25 m thick covering more than 11,000 square km, rich in vertebrate remains from a large synapsid (mammal-like reptile), and having the topology of a perforated sheet (Smith and Kitching 1997). DISC has also yielded remains of large vertebrates (horses and the early camel relative Prosynthetoceras, (Figure 32). They were found at or slightly above its upper surface and probably represent animals buried when deposition of overbank fine sediment resumed in the area. Behrensmeyer (1982) discussed paleosol assemblages representing 4,000-9,000 year durations in her discussion of time interval sampling in a terrestrial environment, and Smith and Kitching (1997) estimated 50,000 years for the accumulation of the widespread condensed bed they studied in the Jurassic of South Africa. The accumulation of the DISC main conglomerate probably took more time than the development of regular paleosol fossil assemblages, because it includes a period of denudation and concentration in addition to paleosol formation, and probably less time than the thicker and more extensive South African deposit.

Bown and Kraus (1981) discussed vertebrate concentrations in mudstones from lower Eocene floodplain sediments of Wyoming, concluding that the bones accumulated as lags in the A horizons of paleosols with little effect of sorting in water. They reported that small mammal teeth make up 70% of the teeth recovered, teeth that would readily travel with sand-sized particles, and used their presence, the lack of sand, and other evidence to conclude that hydraulic sorting was not significant in these deposits (Bown and Kraus 1981 p. 48-49). All of the Fort Polk conglomerates include nodules and bones rounded in water transport (Figure 25-27, and Figure 33), some show distinctive cross-bedding (Figure 24), and some of the most fossiliferous include considerable sand (Figure 27). Many of the fossil remains from the pedogenic nodule conglomerates moved as sedimentary particles when carried by water, as did vertebrate remains discussed by Voorhies (1969), Dodson (1973), and Korth (1979). Those that accumulated in channel lags, for example, the deposit at TVOR Site, Fort Polk, were certainly subjected to such processes. Small mammal teeth (Figure 34) are moderately transportable in water compared to other skeletal elements. They belong to Korth's (1979) dispersal category II or III, with I having the lowest settling rate (ribs), and IV the highest, for example, mandibles. High percentages of fish teeth (including pharyngeal teeth) are found with the terrestrial mammal teeth at Fort Polk (Table 1, Figure 35). When identifiable, they have proved to be from fresh-water fish, that had to be either transported into the areas where nodules and terrestrial vertebrate teeth were accumulating, or the nodules and mammal teeth had to be transported to join them. These factors indicate that the Fort Polk conglomerates include more reworking and transport in water and mixing of materials, and comparatively more time averaging than the fossil accumulations described by Bown and Kraus (1981). The lower part of DISC main conglomerate is more similar to deposits such as those described by Bown and Kraus (1981) than are TVOR or Stonehenge sites, where more transport occurred, at Fort Polk. It shows a lower concentration of all kinds of teeth, fewer fish teeth per unit of weight when compared to mammals, and more large bone fragments versus small bone fragments, than they do (Figure 35, Table 1). A rough correspondence between degree of cementation with calcium carbonate and productivity of teeth, including terrestrial mammal teeth, is seen at Fort Polk sites. Conglomerates like TVOR where more winnowing has taken place have better cementation (Figure 27) and more fossils, including a higher percentage of fish teeth (Table 1). Clayey conglomerates like DISC (Movie 1, Figure 21) have fewer fossils because they represent both less erosion and concentration of the harvest of small teeth from soils, and less transport which added remains such as those of fresh-water fish.

Fort Polk conglomerate sites offer the possibility of studying a progression of sites through time in the Louisiana Miocene. Their fossil faunas are being examined for trends. Lindsay (1972) reported a size increase upsection in the rodent Copemys in the California Barstovian, and initial results examining four sites (from lowest to highest: TVOR, Stonehenge, Gully, and DISC Sites), also suggested a size increase in this animal through time. Further screening and analysis, however, produced a very different picture (Figure 36), emphasizing the value of continuing processing to extract more than the handful of small forms initially available.

The unique situation and method of recovery of fossils at Fort Polk guarantees results as long as the conglomerates can be located and processed. The faunal list, currently including 26 land mammals belonging to nine orders, may eventually include 100 mammals, like the Myers Farm Site in Nebraska (Corner 1976). Although more than 3,000 specimens are catalogued, screening continues because it is especially important to increase samples of new species and rare forms. Hedgehogs (Schiebout 1996) and a new species of tiny beaver are represented by less than ten teeth apiece. In a Miocene possibly warmer than the modern western Louisiana climate, frugivorous bats might have survived, but none have been found as yet.

The close association of vertebrate faunas and pedogenic nodules in the conglomerates offers an opportunity for fruitful geochemical research. Geochemical studies of delta¹³C and delta¹⁸O compositions of pedogenic carbonates from the Fort Polk Miocene paleosols and nodule-rich conglomerates are underway by Paul Aharon (personal commun., 1997; Schiebout 1997b). Pedogenic nodule delta¹³C compositions reflect original soil CO₂ which, in turn, reflects the nature of the original biomass, and delta¹⁸O is known to have a very strong positive correlation with the isotope composition of local rainfall (Cerling 1984).