ARCHEOLOGICAL FIELD RESEARCH METHODS
In April 1992, the NPS submitted a research design on the Palo Alto battlefield to the Texas State Historic Preservation Officer (SHPO). This document (Haecker 1992) provided the framework for a more in-depth survey of the battlefield, as recommended 13 years earlier by Texas A&M (Bond 1979). NPS personnel were to conduct the survey during the summer of 1992. The following is a synopsis of the research design submitted to the Texas SHPO before the 1992 survey. Statements in bold face type are either initial theories or historical assumptions eventually found to be incorrect or debatable, or proposed field methods later modified or abandoned.
Proposed Investigative Theory and Practices
The principal goal of the Palo Alto battlefield investigation is to further define general battle lines identified by previous researchers (Bond 1979; Sanchez 1985; Champion n.d.). Based on their findings, it is presumed:
1) recovered artifacts will be primarily of a military nature consistent with what is known about Mexican and U.S. military items used in 1846;
2) the battle was essentially an artillery duel, therefore, most of the artifacts will be artillery munitions;
3) concentrations of artifacts will generally correlate with battle lines of both sides with U.S. munitions largely on the Mexican line and Mexican munitions largely within the U.S. line;
4) there has been an adverse effect on the site due to relic hunters;
5) some portion of the recovered artifacts may pre- or post-date the 1846 battle, reflecting other historical events;
6) the ground surface is largely obscured due to agricultural and natural vegetation cover. Also, water and wind-related deposition and erosion actions have buried many of the artifacts below the surface, to a depth of approximately 4-6 in; and
7) natural deterioration of the artifacts, especially iron and steel, is great.
Military Equipment and Archeological Expectations
Expected archeological remains associated with the 1846 battle will reflect the nature of the confrontation, that is, an artillery duel. Large-scale, hand-to-hand combat which characterized later battles of the Mexican-American War did not occur at Palo Alto. During this first battle of the war, approximately 3,000 U.S. artillery rounds were fired, while the Mexican artillery fired only around 650 (Brooks 1965:135). A combination of superior cannon, effective use of canister and spherical case shot, and rapidity of cannon fire by U.S. forces resulted in devastation of the Mexican army. Based on various battlefield accounts, most of the artillery-related artifacts probably will be of U.S. origin.
United States Arms and Equipment
To date, there are no known contemporary listings of all types of small arms used by the U.S. army during this battle. However, previous researchers (Bond 1979:18; Bateman 1982:32-36) believe the principal small arms of U.S. troops consisted of the U.S. Model 1835 flintlock musket in .69 caliber, or the U.S. Model 1841 musket or rifle in .52 caliber. Possibly a few of the breech-loading Model 1819 Hall carbine rifles, modified for percussion caps and .52 caliber rounds, also were used. However, an unknown quantity and variety of other types of firearms were present.
Those U.S. troops armed with the Model 1835 musket were issued lead balls in paper cartridges, either in single round-ball and black powder cartridge or a ball with three buckshot ("buck and ball") cartridge. The Model 1819 Hall rifles took a .52 caliber round ball in a paper cartridge. A mix of both flintlock and percussion cap muskets and rifles was in service during this battle; therefore, percussion caps may be found particularly along the U.S. army battle line. Bayonets for the above small arms were of the angular offset-blade socket type with clasp ring (Webster 1965).
No information is available on all the varieties of handguns used in this battle. However, the most common U.S. army-issue pistols present probably included the Model 1836 flintlock and the Model 1842 percussion, both in .54 caliber (Bond 1979:18).
The U.S. artillery in this battle consisted of Model 1840 6-pounder and at least two 18-pounder iron siege or garrison guns. 12-pounder howitzers were not used during the battle. Ammunition for the 6-pounder guns included shot, spherical case and canister; the 18-pounder guns fired shot, spherical case, grape and canister. Shot is a solid, nonexplosive, conventional cannon ball. The U.S. shot was made of cast iron, Mexican shot was commonly made of copper or cupreous alloy. Grape consisted of a stand of nine small iron balls stacked in three rows with three balls to the row and held in vertical column between two iron plates by an iron bolt. Spherical case consisted of a hollow shot filled with lead balls and an exploding charge fused to detonate in flight. Canister consisted of a metal sheet cylinder filled with iron balls (Bond 1979:18).
In 1846, the U.S. Army uniform consisted of dark blue jackets, sky blue pants, and cloth fatigue caps; epaulettes were hardly ever worn in battle (Allie 1991:4). Metal objects associated with this uniform would include gilt brass buttons (with the letter "I", "A" or "D" on the button to indicate Infantry, Artillery or Dragoon, respectively), eagle hat insignias, and belt buckles with the letters "US" (U.S. War Dept. 1841:379-381). Metal objects normally associated with civilian clothing, such as plain brass and lead or pewter buttons, as well as items of a personal nature also may be present along the U.S. battle lines.
Mexican Arms and Equipment
Most of the Mexican small arms were of European manufacture, consisting primarily of the antiquated English India Pattern "Brown Bess" flintlock musket. The Brown Bess was manufactured in .75 caliber and used a socket bayonet (Bateman 1982:38). Other small arms used by the Mexicans may have included the British Baker rifle, caliber .625; the Prussian Model 1839 Tigre rifle, caliber not yet identified; and a blunderbuss-type weapon called an escopette or scuppet, used by the Mexican cavalry and lancers. There probably is no specific caliber for this weapon; rather, any type of lead ball that could roll down the barrel would have been used. Types of Mexican handguns are unknown, but likely only officers and perhaps the cavalry would have carried them. It also is likely such weapons would have been flintlocks (Bateman 1982:39).
Exactly what type of artillery pieces the Mexicans possessed at Palo Alto is not known. However, a list of Mexican ordnance and ordnance stores exists in the inventory of captured property resulting from the battle of Resaca de la Palma, fought the day after the battle at Palo Alto. This list is as follows (Larnard 1846):
Six 6-pounder cannon, field pieces;
Ordnance captured at the later battle of Monterrey, fought in September 1846, showed the Mexican army possessed 3-pounders, 4-pounders and 7-inch howitzers (Brooks 1965). There is no mention of spherical case or grape in the inventories of Mexican equipment from either inventory; however, some battle accounts occasionally mention use of grape by the Mexicans (Brooks 1965; Grant 1885). Most of the Mexican cannon shot was made of copper or cupreous alloys, as opposed to the iron shot used by the U.S. One Mexican copper solid shot was recovered during the 1979 archeological investigation of the Palo Alto battlefield. This artifact weighed 4.76 pounds and measured 3.12 inches in diameter (Bond 1979:19).
No detailed information is available on the uniforms of the Mexican troops that specifically fought at Palo Alto; however, the elite units presumably wore elaborate uniforms that included buttons and other brass accouterments bearing the Mexican eagle and snake. Under a series of detailed regulations issued in 1839, the Mexican enlisted soldiers were to be issued colorful but relatively plain cloth coats and trousers with leather shakos. However, the rank and file reportedly often fought in sandals and cheap cotton garments (Nevin 1978).
Testable Assumptions: Battlefield Locations
Historical documentation and previous archeological investigations of the battle allow for a number of archeologically testable assumptions. These assumptions depend on artifact function and national origin, and a period of use no later than 1846. Keeping in mind the primary goal of the project, the identification of the initial Mexican and U.S. battle positions, the following testable assumptions, primarily from the 1979 archeological results, are as follows:
1) the initial Mexican battle line was anchored on its right (east) flank by a small elevation. From this elevation the Mexican line extended to the west for approximately 2000 yd, up to the now abandoned Matamoros Road; and
2) the initial U.S. battle line was parallel to the Mexican line, and approximately 700-800 yd north of it.
Utilizing what is known about the battle, the initial Mexican line will:
1) contain a much higher level of spent and fragmented artillery ordnance than the initial U.S. battle line;
2) contain cannon balls mostly made of iron;
3) contain solid cannon balls which will be of two sizes: 6-pound (3.5" diameter), and 18-pound (5.2" diameter);
4) contain metallic parts from Mexican uniforms and metal parts from Mexican-related small arms, for example Brown Bess muskets; and
5) contain some copper cannon balls, due to spillage.
In contrast, the U.S. battle line will:
1) contain a relatively lower level of metal;
2) contain, for the most part, cannon balls made of copper; and these cannon balls will be of two sizes: either 6- or 8-pound, and 12-pound;
3) contain some iron cannon balls, due to spillage; and
4) contain a significant amount of musket balls on its west end due to Mexican outflanking attempts there.
Proposed Field Methods
The following proposed field methods were modified in part during the actual survey. Portions of the proposed field methods modified at the time of survey, or found to be incorrect, are noted in boldface type and are discussed in Chapters Seven and Eight.
The 1992 field investigation will consist of three phases: orientation, survey and recording with collection. There will be no discrimination in recording and collecting on presumed period association. Aboriginal, pre-battle, battle-related, post-battle related and recent historical artifacts will be considered equally.
Aerial photographs will be made of the battlefield before archeological fieldwork begins. The area to be flown measures 14,200 ft North-South by 11,200 ft East-West, encompassing approximately 3,670 acres. The actual battle area to be surveyed is well within this acreage. Nominal scale of the photographs will be 1:6000 (1" = 500'). Paneled ground control points, established using the taking of the aerial photographs, will provide the control needed to allow for photogrammetric mapping, employing analytical bridging techniques.
The surveyors who establish the paneled control points also will establish a series of staked data points, spaced 1,000 ft apart, between the paneled control points. These additional data points will provide a reference grid for the survey, recovery and recording phases. The archeology field crew will establish additional staked data points as needed and tied into the reference grid. All data points will be tied into the U.S. Geological Survey bench mark presently located on the battlefield.
The project area will be surveyed based on a system of prioritized parcels and tracts, using information from the methodological approaches and findings of the 1979 season. Parcels and tracts given the highest priority are those not archeologically surveyed, or minimally surveyed in 1979, that possess a high probability of containing battle-related artifacts. Second priority areas will be those portions systematically surveyed in 1979 and found to contain battle-related artifacts. The lowest priority will be given to those areas that, according to historical and archeological documentation, are outside the battlefield proper and, therefore, outside the present scope of work. However, these low priority areas will be included in future investigations since they may contain evidence of the earthworks that protected the U.S. army's wagon train as well as evidence of the Mexican army's post-battle mass graves and encampment. To identify their presence, such features will require other methods of remote sensing not scheduled for this proposed project.
All non-tree vegetation within the above-selected survey areas will be removed before the actual survey. This must be done to allow the use of metal detectors as well as for visual inspection of the ground surface. The chosen method of vegetation removal, either by close cropped mowing or by burning, will be based on the preferences of the various parcel owners.
The chosen method of survey is the use of electronic metal detectors. Visual inspection of the surface will be done, when possible, concurrently with the metal detector survey. The survey crew will consist of two professional NPS archeologists and local amateur volunteers, with the volunteers supervised by the archeologists. All personnel using metal detectors will be aligned at 15-ft intervals and oriented to grid cardinal directions. The surveyors will proceed in line, using a sweeping motion to examine the ground, with each operator covering a sweep of approximately 5-7 ft. A pin flag will be placed at each target located by an operator. As soon as the location is pinned, the operator will continue along the transect. In some instances, the target location will be exposed immediately so the operator can check on machine performance, that is, establish the degree of accuracy of a given machine in ascertaining artifact depth, metallic type and size. Objects found on the surface also will be marked with pin flags and left in place for recording and recovery.
Recording and Collection
Individual artifact locations will be measured to within 1-in accuracy in distance using an electronic theodolite. Provenience data will then be downloaded into a PC software program and processed at the Division of Anthropology, Branch of Cultural Research, Remote Sensing Section, Southwest Regional Office of the NPS. This will create a computer map, showing artifact distribution. Since this data is tied into the aerial photos of the site via ground controls, it will be possible to overlay artifact distribution onto a photogrammetric map.
The collection phase will consist of the removal of both surface and subsurface artifacts. As noted in Bond (1979), subsurface artifacts are within 6 in of the ground surface. Artifact removal will be done as soon as it is documented, as it is possible pin flagged artifacts left overnight could be taken by relic hunters. Perishable artifacts, if encountered, also will be recorded and collected for analysis and curation. In the unlikely event human remains are encountered, no further exposure will occur after confirmation. Such remains, as exposed, will be described, photographed, and their exact provenience obtained. Any associated artifacts will be provenienced, photographed in situ with the remains, and collected. Exposure of human remains will be limited as much as possible. Information regarding such a discovery will be limited to NPS personnel for protection of the remains.
Following NPS Archeological Documentation Manual guidelines, field records will be kept in a field specimen (FS) catalog, with each recovered artifact given a sequential FS number. A general description of each artifact at the time of its recovery also will be included in the FS catalog. Each artifact will be placed in its own container, appropriate for its size and state of fragility.
The site will be photographed with both black-and-white print and color slide film. Photographs will document the general physiographic setting of the site as well as the various survey activities. Where appropriate, each photo will contain a scale, north arrow and sign board. An in-field photo log will document each exposure and include the subject, its FS number, direction of view, date and photographer. A video camera also will be used to document the project.
Fieldwork also will include the recording of artifacts previously collected by relic hunters as well as those now curated by local museums and archeology laboratories, for example, Texas A&M and TI&M. Provenience information for artifacts in the possession of relic hunters will be general at best; therefore, such information only will be used to supplement the findings of this investigation. NPS personnel will obtain descriptive documentation of those artifacts in the possession of cooperative relic hunters. If possible, the NPS will acquire, either by gift or loan, these artifacts for further study and possibly for display at the proposed Palo Alto Battlefield Visitor Center.
State-of-the-art laboratory procedures will be used in cleaning and stabilizing the artifacts. The actual task will be contracted to a professional conservation laboratory. Essentially, the tasks will include the removal of accumulated dirt from each artifact, then determining the condition of the artifact to ascertain what conservation method(s), if any, is needed. Sorting and identifying the artifacts will be undertaken by personnel experienced with artifacts of this period. More specific artifact analysis, involving the subsorting of artifacts into further identifiable discrete types, will be conducted by the NPS Project Director and a trained assistant.
Artifact analysis will address the basic goal of the research orientation which is the identification of the Mexican and U.S. battle lines. Categories of investigation will include: artifact function such as arms and ammunition, uniforms and personal possessions; and diagnostic attributes such as manufacturers' marks and military unit insignia. Occurrence frequencies for artifact types, for example, Mexican and U.S. uniform buttons, iron and copper artillery ordnance, .69 caliber musket balls, in conjunction with their provenience data, will result in the identification of the location and nationality of battle lines. The artifacts, notes, records and other documentation will be archived at the NPS Southwest Regional Office in Santa Fe, New Mexico.
Applied Field Methods of the 1992, 1993 Field Seasons
Before beginning the 1992 survey, a permanent reference grid was created over the proposed 3,400 acre park. The grid consisted of 50 permanent data points spaced 1,600 ft apart, not 1000 ft as was proposed in the research design. This modification was possible due to the ability of the theodolite to accurately read distances of up to .5 miles. Four temporary data points were later added, placed within the identified core battlefield area to facilitate proveniencing of artifacts. A permanent datum consisted of a piece of steel reinforcing bar pounded flush with the ground with an aluminum cap giving its numerical designation. For convenience the grid was oriented parallel and perpendicular with Farm-Market Road 1847, which is the western boundary of the proposed park; this road is aligned on grid north.
Orientation also included teaching volunteers about battle history, proposed survey methods, artifact identification and recording methods. Several of the volunteers are experts in the use of metal detectors; they provided instruction regarding proper operation of these tools.
As is often the case for an archeological project, a disparity in field techniques occurred, regarding what was proposed and what reality required. The 1979 survey report recommended 100 percent removal of the vegetation before any future survey of the battlefield. This recommendation was based on the earlier surveyors' observation that mowing the cordgrass left stubble that prevented adequate contact of the metal detector coils with the ground surface (Bond 1979:27, 58-60). Since mowing the cordgrass apparently was an inappropriate method of vegetation removal, burning off the vegetation was the logical alternative; a burn-off of a portion of the battlefield by a landowner in 1990 exposed enough ground surface for the successful use of metal detectors by relic hunters (Walter Plitt, personal communication 1992).
However, several environmental and legal concerns quickly became issues while scheduling a controlled vegetation burn-off of approximately 600 acres of cordgrass: such a burn-off would have an adverse effect on potential habitat for a number of rare and endangered species that could exist in the area; the resulting smoke of such an extensive fire would be a major source of air pollution; and the fire could escape its confines and damage adjacent lands not authorized for burning.
Fortunately, while the above environmental concerns were being addressed, it was possible to survey portions of the battlefield not requiring vegetation removal. The first area scheduled for investigation was an 80 acre plowed field, chosen for its absence of vegetation (Figure 36). In 1979, this field was subjected to a systematic magnetometer and metal detector survey, resulting in the location of only five battle-related artifacts (Bond 1979). It was initially presumed the results in 1992 would be similar or worse. However, for the purpose of learning how to use the metal detectors, any subsurface metal artifacts would suffice.
The initial results were unexpectedly rewarding. After three days of survey, 26 battle-related artifacts were recovered within three sample areas, totalling seven acres. In other words, instead of recovering at best one battle-related artifact per 16 acres in this particular field as was done in 1979, one could now expect close to four such artifacts per acre. This notable improvement in surveyed acreage/artifact recovery ratio was attributed to improvements in metal detector technology made since the last archeological survey of Palo Alto. Familiarity with various metal detectors used during the initial phase of survey indicated the most technologically advanced models are capable of: giving accurate readings of subsurface metal even when the metal detector is suspended over four inches above the ground surface; discriminating between metal and salt concentrations that often were responsible for false readings by earlier models; and providing readings of metal to a depth much greater than the metal detectors used during the 1979 survey.
These observations provided an alternative to burning off the cordgrass, which was close-cropped mowing as was done in 1979. In addition, the use of superior metal detectors permitted survey of densely vegetated areas that, although not covered by cordgrass, can still inhibit the effectiveness of other, less sophisticated metal detector models (Figure 37).
It was now obvious the battlefield contained many more artifacts than initially assumed (Figure 38). It also was apparent a 100 percent survey of hundreds of acres was now beyond the scope of the 1992 field season; at the rate of three surveyed acres per day, it would take over two years to complete the survey of the core battle area.
Accordingly, several sampling strategies were devised in place of prioritizing tracts for intensive, 100 percent survey as was initially planned (Figure 39, pocket insert). These strategies consisted of:
1) contiguous placement of seven 500 x 100 ft Search Areas (SAs) along a north-south axis, forming one transect measuring 3,500 x 100 ft. This transect, consisting of SAs 3-11, was positioned to find the approximate western edge of the core battle area. Eight other transects, each measuring 3500 x 30 ft along a north-south axis, also were surveyed. These transects were positioned in series to find the approximate eastern edge of the core battle area (Figure 40). It was expected survey of these transects would result in discovery of segments of both Mexican and U.S. battle lines;
2) discrete placement of 16 SAs, each measuring 500 x 100 ft. and all within the core battle area. Placement of these Search Areas was done using a stratified random sampling strategy, involving random placement of sample SAs within historically documented activity localities of the core battle area; and
3) reconnaissance-level survey sweeps, involving the use of one or more metal detectors over informally demarcated or topographically defined areas, such as pastures and resaca levees. This survey approach could quickly determine the presence or absence of artifacts within a given area, occasionally providing guidance on the placement of survey units and indicating areas for future site investigations. Artifacts discovered via reconnaissance were recorded and their locations plotted, then reburied (Figure 41, pocket insert).
It also became obvious too much of the limited survey time would be concentrated on recording and collecting items of obvious recent manufacture and deposition, for example, aluminum cans and fence staples. Accordingly, the project director made the decision regarding what to record or not record.
Placement of transects, survey units and reconnaissance sweeps involved: 1) a general knowledge of battle lines, as per eyewitness battle narratives, historic battle maps and provenience of previous artifact findings, thus maximizing the probability of finding battle-related artifacts; 2) surveying only those properties where landowners gave their written permission; and 3) avoidance of potential faunal habitats or rare and endangered species, as defined and proscribed by the U.S. Fish and Wildlife Service.
The four activity localities sampled during the 1992 field season include:
1) the presumed scene of action where both the U.S. and Mexican cavalry units attempted flanking maneuvers on the western side of the core battle area. Both sides failed in this tactic primarily due to the successful defensive use of cannon. This locale should contain quantities of iron and copper grapeshot as well as some cavalry related equipment and small arms ordnance. SAs 1-12, 99-101 were positioned within this activity locale;
2) the initial battle position for the Mexican left flank and pre-attack staging area for the above-mentioned Mexican cavalry units. This locale should be situated approximately .5 mi south of the U.S. Fifth Infantry position, according to eyewitness accounts. Expected artifacts include Mexican cavalry and infantry-related equipment intermixed with U.S. cannon ordnance. SAs 200-208 were positioned within this activity locale;
3) the area north (Figure 40) and east of "Arista Hill", a heavily vegetated low rise. Local historians hypothesize this topographic feature is the low rise that, according to battle accounts, anchored the initial Mexican right flank. If this hypothesis is correct, artifacts found near "Arista Hill" primarily should include Mexican cavalry and infantry-related equipment, intermixed with U.S. ordnance. The presence of a resaca adjacent to this low rise also offers the possibility of locating Mexican soldiers' graves, since the soil within such a marshy area would be easier to excavate for burials than the surrounding floodplain. SAs 300 and 301 were positioned east and northeast of this rise, with SA 302 placed along its northern flank; and
4) the vicinity where relic collectors reportedly found significant concentrations of both copper and iron cannon balls, and articles related to the U.S. army. According to Palo Alto Committee Chairman and local historian Walter Plitt (personal communication, 1992), this activity locale is within the present-day transition zone of cordgrass and mesquite, just south of an east-west dirt road and east of a north-south dirt road (Figures 3, 39 and 41 pocket inserts). Such an artifact concentration might roughly correspond with the U.S. battle line and perhaps placement of one of its batteries. Relatively large numbers of cannon balls here could reflect Mexican attempts to destroy U.S. cannon.
It was expected the north ends of SAs 306-313 would sufficiently address this area of known artifact concentration. Therefore, SAs 304 and 305 were placed to the immediate north of the east-west dirt road and east of the north-south road, to sample an area presumably not collected by relic hunters due to its thick vegetation cover.
Using the above field methods, archeologists and volunteers together surveyed 54 acres during July-August, 1992. This encompassed approximately 4.5 percent of the estimated 1,200-acre core battle area and 1.5 percent of the proposed 3,400-acre Palo Alto Battlefield Historic Site.
Field methods used during the 1993 field season were essentially the same as for the 1992 field season, with a notable shift in emphasis of sampling approaches. The reconnaissance sweep survey method was used to a greater extent, with correspondingly less emphasis on selection of stratified random survey units slated for 100 percent artifact recovery. This shift in survey methods was based on an observation made toward the close of the 1992 season, that is, the presence or absence of subsurface artifact clusters was best determined by first conducting a series of widely spaced metal detector transect sweeps over an informally defined target area, for example, an 80 acre plowed field. Such reconnaissance-level sweeps, representing an approximate five percent sample of a given target area, could be completed by a few metal detector operators in one or two days. Using this approach, an artifact cluster, if such existed within the target area, initially would appear as a widely spaced distribution of only a few artifacts.
The next task was to further delineate the artifact cluster. This was done via nonstructured but intensive sweeps around each of the individual artifact locations. These intensive sweeps would determine if an artifact in question was essentially an isolated occurrence or one of a number of artifacts within a cluster. If an artifact cluster existed, more artifacts would soon be discovered. Conversely, if no other artifacts were found nearby, the artifact would be labeled an isolated occurrence. Thus, information on the presence or absence of artifact clusters within a target area could be obtained using minimal time and personnel.
The above-described survey method has its limitations in that it cannot provide data allowing for relative comparisons of artifact cluster densities, nor can it provide a comprehensive list of artifact type(s) within a given cluster. Such data can be gained, however, through intensive metal detector sweeps of formally defined sample units. Four such sample units, SAs 209-212, were surveyed for total artifact recovery during the 1993 field season. These sample units were specifically placed to further delineate the Mexican battle line discovered during the 1992 field season. Only one artifact cluster of a U.S. battle line position was ever discovered during both field seasons, the widespread artifact cluster defined by SA 305 and the extreme north ends of SAs 306-308. As a result, during the 1993 field season, one intensive sample unit survey was deemed appropriate towards defining a portion of the U.S. battle line.
During the 1993 survey, 15 acres were intensively surveyed; an additional 210 acres were subjected to a five percent sample sweep. Thus, both field seasons combined resulted in an estimated eight percent sample of a 3,400 acre area.
Last Updated: 25-Feb-2009