software known as Access. This was an intimidating and time consuming task. We were however ultimately successful. It was at this point we began to use the Windows operating system which is a much more user-friendly operating system. Again we upgraded and added more memory and new fields to our database.

Bill’s knowledge of shipwreck scatter from the Atocha days was immense and we began to discuss a method of entering all of the artifact locations from the past into our database so that we could expedite the plotting of everything on one computerized chart. The hope was that this would enable us to make more intelligent decisions about where to concentrate our search. Unfortunately, because we were creating hand drawn charts and the data we had been entering up to that point was intended for archival purposes, it was limited to the original fields and did not include location positions. This necessitated going back to the beginning and entering all of the location data.

The next step was to establish the location of the beach markers from which the salvage vessels took their positions. At that time the salvage vessels were using sextants to take two angle positions on three beach markers that were on land. The beach markers had been surveyed by State officials (with the assistance of salvors) using the DNR (Department of Natural Resources) benchmarks for their origination points. The first "surveyed" beach markers were put up in 1983. Prior to that time the salvors used many things for positioning, mostly landmarks such as piers, trees, houses etc. Few of these positions are repeatable today because houses are painted, modified or moved, trees fall or are chopped down, piers wash away and more homes and condos are built, etc. Therefore, the majority of the pre-1983 data is not able to be plotted. After the beach markers were placed they were subject to vandalism, winter storm surges, souvenir hunters, and sometimes even adjacent landowners, who thought that the markers blocked the view of the ocean from their million dollar beach homes. At the beginning of every season, (and sometimes mid-season) the beach markers were replaced, and not necessarily in the same position. All of the notes of how and where these beach markers were placed were handwritten in a field notebooks and filed in the state files in Tallahassee, Florida. Most of the notes were legible, some were not and some were even lost. With the help of our then current archaeologist, Catherine Gaither, we were able to calculate the positions of most of the beach markers on the various sites over the years. We then had to develop a trigonometric formula

that the computer could understand to calculate the locations of the artifacts based on the triangulation readings and translate them from UTM numbers into State Plane x/y’s which was the format all of the government maps were in. (See Florida East Coast Shipwreck Project, 1997 Season Report, Pgs. 37-39)

We solicited help from outside consultants and volunteers. After more than a year of research and development, the combined efforts of volunteer Charlie Myer, our conservator Bill Moore, archaeologist Catherine Gaither, subcontractor/salvor Ken Nehiley, and myself, a program was produced in Microsoft Excel- a spreadsheet program, which could make the conversions. We used the Douglas Beach site as our prototype. Each of the wreck sites had a different numbering or lettering system for beach markers, sometimes ascending and sometimes descending. We had to reconfigure the program for each site and for each year to make all of the conversions. We tested a state plane map from the State of Florida on disk and many other mapping programs that were currently on the market. One of the problems we encountered was that most of these programs were able to plot only one "find" location at a time. This was a problem since we had accumulated over 20,000 items to plot in our database, so this method was not feasible. We also discovered that most mapping programs did not allow one to zoom in to an area as close as we needed to see it. Additionally, they had a fixed size for their plots which was much larger than we wanted. The majority of people who use these programs do so for navigation and do not require the accuracy that we do. We came to the conclusion that no one in the world except us cared if they were within a few feet of where they wanted to be.

Latitude and longitude is a familiar co-ordinate system while the UTM (Universal Transverse Mercator) and/or State Plane co-ordinate system is not. Eventually, in 1995, Bill Moore figured out how to insert the State Planes map into AutoCAD Lite (an engineering program) and plot the artifact locations using Microsoft Word to create a mail merge that took data from the Microsoft Access database and create a file that could be read and plotted in AutoCAD Lite and make it look like a real map. Later, Bill designed a lat/long map on which we could plot that the salvors could use. The archive was becoming extremely large and the RAM (random access memory) requirements of manipulating all of this data made it necessary for us to upgrade our computer to a 386 with an expanding memory capability. Backing up this data was tedious and took over thirty 3.5" disks (1.43

megabytes each) so we also obtained the new "Zip" drive with "Zip disks" with 100 megabytes of storage capability apiece.

One of the largest problems we faced was relocating beach markers every year and then recalculating all finds accordingly. Human error and Mother Nature combined to make it seem almost insurmountable. At this point we decided to ask the State of Florida to allow us the use the Global Positioning System (GPS) instead of sextants to establish the artifact positions. We spent two years testing the GPS system and writing reports (See Florida East Coast Shipwreck Project, 1995, 1996, 1997 Season Reports) to convince the State that GPS was as accurate as (or more accurate than) the sextants and allowed for less human error. Finally, the state agreed that Differential Global Positioning (DGPS) would be allowed as an alternative to sextants and beach markers. This saved the State money and time as they no longer had to maintain the beach markers. A DGPS system cost a little bit more than a sextant at the time, but it was money well spent. By 1997 all of our vessels and all of our subcontractor’s vessels were using only DGPS. During the next year, data entry continued. Plot files were built upon and improved. Plots were colorized to indicate artifact composition. We obtained a used plotter (giant printer) to produce maps electronically. Scatter patterns never noticed before were becoming clear. The database fields were expanded to include every note on the field note forms including but not limited to date of recovery, water depth, etc. It was about this time that hindsight set in. We had not entered the locations of the empty excavations, only those of actual finds. This meant that we would conduct excavations based only on knowledge of actual finds, thereby revisiting empty holes over and over again. We had to go back to square one and began to enter all the locations of 17 years of empty holes.

In 1998 and 1999 we started examining the feasibility of placing aerial photographs behind the plots to show the shorelines and rock outcrops, shoals, sandbars, etc. to better understand the wrecking process. Aerial photographs must be taken on just the right day at the right time to obtain the best water visibility. We found through inquiries that this is an expensive process. We searched many places and found many aerial photographs which had already been produced, however, most of them had not been done using a system which made it possible to reference the photos into exact locations on our maps. (Geo referencing). Finally however, we obtained aerial photographs of the coast of Indian River county that covered many of the lease areas. Unfortunately, we still have not obtained any

usable photos of St. Lucie or Brevard counties where 5 of our sites exist. The Indian River county photos were originally done for a beach re-nourishment project. These were a composite of aerial photographs (raster images) that were pieced together then split up into thirty-two individual digitized photographs in a tiff format. All the digitized photographs were geo-referenced to the NAD 27 and NAD 83 State Plane coordinate systems. Although geo-references were provided by the county along with the photographs, the geo-references were for the NorthWest corners of the photographs. This created a problem because the customized AutoCAD mapping system we utilize automatically inserts photos with the beginning point on the SouthWest corner of the image. It was therefore necessary to use the latitude of one photograph with longitude of the next photograph to "correctly overlay the image in the proper position".

Using CORPSCON translation software which we obtained from the county property appraisers office, the State Plane coordinate systems were translated into latitude/longitude mapping coordinate systems we used for plotting the artifact locations. John Tokarz, a subcontractor, assisted in the more complex translation of the State Plane system photographs into our charts that displayed our sites in latitude and longitude. This involved sizing the aerial photo in the latitude and longitude dimensions then rotating the photos to correct for northern variation between the coordinate systems. All the photos were then superimposed over the AutoCAD plots of the east coast wreck sites. Once the photos were inserted, our computers started to act like snails. Computerized photographs are huge files which take up a lot of memory space and slow down your ability to process and manipulate data. To continue with the project it became necessary to obtain the new 20 Gigabyte computer with a c/d rom writer (with a storage capability of 600 megabytes per c/d) for backing up the data and storing photos.

This AutoCAD mapping system contains layers which be can turned off or on so that one can view the entire map, or just portions of it. For example, one could turn off all layers except ceramics and view only the scatter pattern of the ceramic artifacts. One of the layers is a rough outline showing the main land features of the coast of the Atlantic seashore and the intercoastal waterway.

Using DGPS fixes on predominant landmarks and satellite photographs of the Florida east coast, the final AutoCAD maps were refined to minimize any errors that were introduced by distortions in the aerial

photograph and the inaccuracies in the coordinate translation process. A latitude/longitude grid system was then overlaid over the entire length of our east coast automated mapping system. This grid system was composed of rectangles representing one- tenth of a minute. This lat/long grid system can be further defined by importing an even smaller grid that overlays the area of interest with rectangles representing one hundredths of a degree spacing. These new charts with the aerial photograph overlays show good details of the ever-changing underwater sand, shoals, rocks and worm rock outcrop areas along the coast and should act as a valuable aid in locating possible areas for exploration.

Our printing limitations at this time make it unfeasible to print out chart sized photos showing all of the plots. To print this large photo would take about 8 hours and 8 ink cartridges and the image is not as clear as it is on the computer monitor. We do make 8X10 print outs with the photos for reference. (See Figure 1) However, the plotter is capable of quickly printing the thousands of excavations in several colors on a large chart format (36X48"). When we first started integrating the empty hole data onto the site maps, many of the artifact finds were not visible because the plotted empty holes covered them up. We then created a new field in the database called "plot priority" which prioritizes the order of the artifacts when plotting so that the least important items (empty holes and modern intrusions) plot first and then shipwreck material plots on top. Another system implemented was to change the color of the empty excavation layer to a light gray so that one can still see all empty excavations on a plot, however, they are easily discernible from actual artifact locations.

We have also found a way to include in our plot files more data from our database so that at any given time one can zoom into a plot on the monitor and click on "list attributes" to get details about that specific plot (i.e. water depth, actual lat/long/ tag number and description of artifact, date of recovery, vessel which made the recovery etc.). This information enables us, within seconds, to access photos and drawings of the artifacts that are embedded in our database.

In the year 2000, we upgraded our computer system to a full and more powerful version of AutoCAD 2000 and Microsoft Office 2000. Our