The documents in this collection are reproduced in two ways: first, as a searchable text and, second, as a set of facsimile (page) images. The facsimile images of the documents provide the most authoritative texts. The transcriptions (searchable texts) are more than 99 percent accurate, but careful researchers will always compare the transcriptions to the facsimiles of the original documents.
Transcription and coding.
In the transcriptions, capitalization and spelling reflect that of the original document. Where the text was illegible, bracketed question marks represent the approximate number of words that cannot be interpreted; e.g., "[???] while he [??]" means "[three unintelligible words] while he [two unintelligible words]." Where a good guess could be made, the word and a question mark have been placed in brackets, e.g., [malversation?].
When initially transcribed, these texts were marked up in Standard Generalized Markup Language (SGML). The American Memory SGML markup scheme conforms to the guidelines of the Text Encoding Inititiative (TEI), the work of a consortium of scholarly institutions. Since this Internet presentation employs the conventions of the World Wide Web, the SGML markup has been simplified and reprocessed to create documents in HyperText Markup Language (HTML). In the final version, SGML markup will be utilized.
Interested persons may obtain the American Memory SGML document type definition (DTD) and related information by file transfer protocol (ftp) from the Library of Congress server. [Full address to be supplied when ready.]
Most of the 22,500 facsimile page images in this collection carry TIFF headers and employ CCITT Group 3 compression. The following images are exceptions; they also carry a TIFF header, but employ CCITT Group 4 compression.