Accessing paleontological specimens curated at facilities throughout the world increases both the cost and time required for research in evolutionary biology. Although a number of electronic databases currently name the facility where specimens and/or assemblages of North American vertebrate collections are housed (e.g. FAUNMAP)* no existing database yet contains detailed descriptive information on individual specimens. Published data for fossil specimens varies greatly from specimen to specimen. depending on the nature of the publication and the research interests of the investigator. Other researchers addressing different issues cannot necessarily find published data that is comparable to their own data and specimens. The proposed effort will address these problems in biological research by developing: methods for collecting or compiling digital representations of paleontological samples. database methods for efficiently storing and cataloging these digital representations a search-engine for locating specimens in a distributed network of digital archives. and interactive visualization and measurement of these digital representations across the internet. A variety of techniques exists for creating digital representations of physical objects that could allow simultaneous access and interactive measurements of specimens from collections throughout the World. Such representations may range from simple photography with selected perspective views to full volumetric representations created by a medical imaging CT scanner. Each method will have unique characteristics in terms of the nature and accuracy of their spatial representation. as well as cost. The proposed effort will develop standardized digitizing methods for paleontological samples and appropriate models for estimating the uncertainty of measurements derived from these various types of digital representations. Efficient access to these digital representations will require the appli cation of database methods that cover the traditional needs for cataloging physical specimens as well as capturing relevant aspects of the methods by which the sample was digitized. This effort will emphasize the development of a scalable architecture that will allow the implementation to vary among a number of distributed institutions, in order to accommodate their available resources. These distributed databases will be accessed through an intemet search engine front-end. Specimens in these databases can then be selected and object-oriented programming methods will be used to provide interactive visualization and measurements that are appropriate for the given type of digital representation. Interactive measurements will be supplemented with estimates of spatial uncertainty derived from models of geometric error. These models will be parameterized by database records describing the methods and circumstances of data capture for each item. Abslxact July 25, 1998 1 This system will be developed in direct collaboration with paleontological researchers and will be developed and tested using dental specimens of various species of mammoth from Anzo Borrego National Park and the Burke Museum at the University of Washington. Seattle. The ability to rapidly relate newly recovered specimens to objective, exhaustive and publicly accessible digital representations from archives around the wo~d would support a vast array of scientific investigations in biology, as well as other disciplines like archeology. In addition. this capability will prove invaluable to museums, libraries, and K= 12 education. Abstract July 25, 1998 2
