METHODS

The first requirement for standardized surface sampling is to have a clear research goal in mind, because this affects the type of data recorded and the deployment of people doing the recording. In the case of the Siwalik surveys, we were interested in delimiting a number of biostratigraphic events that were suggested by more traditional survey and collection methods, i.e., the level of disappearance (LAD) of Sivapithecus, the decline and extinction of the Tragulidae, the regional appearance datum for “Hipparion” and a large genus of bovid, and the overall faunal response to changes in fluvial systems and climate between ~8.0 Ma and 6.0 Ma. Thus, we designed our sampling strategies to cover sequences of exposures in several different areas that fell within this time range, with correlations between areas based on the well-documented Siwalik magnetostratigraphic record (Tauxe and Opdyke 1982; Johnson et al. 1985; Barry et al. 2002). We were also interested in assessing fossil productivity of different stratigraphic levels throughout the Potwar sequence, as we knew from more traditional collecting that it was likely to vary considerably and was undoubtedly important in considerations of biostratigraphy of these intervals. We thus broadened sampling to include fragmentary fossil debris (“scrap”), even if unidentifiable except as vertebrate remains.

In addition to providing data relating to our primary goals, the biostratigraphic surveys also can address questions such as: 1) the effect of different outcrop slopes and lighting (bright sun, overcast, etc.) on fossil collecting, 2) variation among different individuals in finding fossils, 3) variation in skeletal parts preserved in different lithofacies and stratigraphic intervals, 4) ratios of “good” fossils (i.e., identifiable to major group, etc.) to scrap, 5) variation through time and by facies of aquatic versus non-aquatic vertebrates, and 6) relative frequency of recovery of small mammals on walking surface surveys (as opposed to crawling the outcrops). We do not attempt to treat all of these questions here but point them out as possibilities for future research.

As erosion proceeds along the Potwar Plateau strike valleys, resistant lags of carbonate nodules, gravels, and fossil bones tend to be dispersed widely over the ground surface. Occasionally there are patches of more abundant fossils weathering out from particular lithofacies, and these are treated separately as localities (Barry et al. 1980; Behrensmeyer and Raza 1984). The biostratigraphic surveys target the scatter of fossils between the richer patches, although many of the more evenly dispersed remains may ultimately have been derived from spatially circumscribed concentrations. Relatively few Siwalik fossils are found in situ, and the fragmentary remains recorded on the surveys represent the net result of original (pre-fossilization) taphonomic processes combined with erosion and fragmentation on the modern outcrop surfaces. Nevertheless, there is little chance that fossils from higher or lower levels contaminate the level being sampled, given the continuous strike ridges that separate these levels (Figure 3). The areas used for the biostratigraphic surveys either had not been previously searched or were documented only on the basis of spatially circumscribed fossil localities.

Survey parties usually consist of three to six individuals, each of whom is given a biostratigraphic survey card to fill out (Figure 4) as he or she covers the assigned area of outcrops. Well-defined blocks of exposures are typically surveyed by several individuals together, evening out differences in experience between individuals. One individual is responsible for assigning the search area and for keeping people on more-or-less parallel tracks along the outcrops, for noting light and substrate/slope conditions, outlining the area on airphoto overlays, and collecting the cards at the end of each survey. Individuals vary in their ability to identify fossils, so the more experienced typically assisted the less experienced. In practice, most of the surface fossils can only be identified to major vertebrate or mammal group, which is relatively easy even for inexperienced collectors.

For each biostratigraphic survey block, the surveying team typically spends several hours walking along a dissected, low-relief strike valley or proceeding carefully across the steeper slopes below a capping sandstone looking for fossils. When a team member finds a fossil, decisions must be made during the recording process about the identity and size of the bone or tooth fragment (Figure 4). A number of rules were developed to standardize recording and collecting of some of the materials encountered on the surveys.

1. At the start of each biostratigraphic survey block, the surveyor fills out the top portion of the card, and at the end of the block notes the finish time and total hours. Individuals were asked to stop the clock during any significant rest-breaks during the surveys.
2. All bones are recorded, either tallied as “scrap” if unidentifiable or “turtle scrap” if identifiable as such, or as separate items numbered sequentially and identified to body part and taxon. Body part can be as non-specific as “limb fragment - mammal,” or as specific as “upper right premolar - Hipparion.”
3. Patches of multiple fragments of bones or teeth that have recently broken up on the eroded surfaces are counted as single occurrences.
4. Each bone is scored as either larger or smaller than 5 cm maximum dimension, to keep track of the degree of fragmentation in a survey block and also the observation capabilities of different surveyors.
5. Two or more identifiable and separately recorded bones from the same or different individuals that occur in a small area (e.g., ~1-10 square meters) are noted as “clusters” by brackets on the survey card (Figure 4). Some of these were later designated as localities and given a locality number.
6. If a surveyor is stuck on fossil identification, he/she calls in another surveyor for a second opinion, or in some cases collects the specimen for later identification. We often convened periodically to go over identifications, and in practice, most people kept close enough together that it was easy to check an identification without stopping the survey.
7. The group leader(s) are called in to collect particularly good or informative specimens, which are documented on air photos or, in some cases, simply to survey level and block. Surveyors are instructed to leave such specimens in place and to put a cairn at the discovery site, then get the assistance of other team members. If air photographs and/or a GPS are available, map position and GIS coordinates should also be recorded.
8. At a particularly rich patch of fossils, i.e., a locality, the survey clock is stopped and the team gathers to collect the patch and document it as a locality. Once this is done, the clock starts again, and the survey continues.

This procedure can result in a lot of information in a relatively short period of time, depending of course on the density of surface fossils and the size of the survey team. It can also result in the discovery of high-quality specimens (Figure 5) as well as the collection of additional identifiable material to supplement formal localities (Barry et al. 1980, 2002). The standardized methodology allows various kinds of analyses that are not possible with more free-form paleontological surveying. In the examples to follow, we focus on measures of fossil productivity – identifiable bones and teeth per search hour, ratios of different taxa or body parts – e.g., bovid versus equid and teeth versus axial remains, abundance of a particular taxon relative to total identifiable sample, etc. The fossil productivity measures can be used in conjunction with geological data to investigate how sedimentary environments or stratigraphic intervals vary in fossil richness. Taxonomic data, even at very coarse levels of identification (e.g., mammal versus reptile) can provide evidence for questions such as the abundance of aquatic components in the fauna. The value of such information is most apparent when the data are compared through time or across different areas representing the same time.