218 University of California Publications in History
edition. It shows Louisiana as extending only to the Guadalupe
River (a stream slightly northeast of the Nueees River, the western
limit of the old Spanish province of Texas), but the later editions
all employ the Rio Grande as the boundary. This change might
have been due to Thomas Jefferson, to whom Melish sent a copy
of the map in 1816, and who wrote a most interesting letter in
reply.' Jefferson was sure that the Rio Grande should be the bound-
ary, but at this time seemed less confident than he was at an earlier
date about our claim to the Oregon country. He thought we had
a title, though, in Gray's discovery and Astor's settlement.
 Section two of the Geographical Description included discus-
sions of the British and Spanish boundaries and of the major
rivers.' In it the history of the French exploration of Texas and
"A Map of Louisiana and of the River Mississippi," published in
John Senex's New General Atlas (London, 1721), were cited as
authorities for showing Texas as a part of the Louisiana Purchase.
For the far Northwest, Melish admitted having no positive grounds
for his delineation, but said he believed the title of the United
States to the Columbia watershed "unquestionable." He said as the
Spaniards had not occupied the region north of San Francisco it
might "be considered virtually a part of the United States ter-
ritory, provided they should consider it of importance to take
possession and settle it." He recommended such occupation. The
supposed river San Francisco, flowing from what is now Utah
into San Francisco Bay, he said, "presents itself as a convenient
boundary between the United States and Spanish settlements."
 Most of the major inaccuracies in the western portion of the
 map are common to all those of the period. Two deserve mention
 as giving rise to confused controversies in the Adams-Onis nego-
 tiations. One was in showing the Multnomah (now the Willamette)
 River, which Onis proposed to use in the demarcation, as being a
 great deal longer than it actually is. The other was the uncertainty
 concerning the location of the source of the Arkansas, a specific
 point used in the boundary definition, which no doubt suggested
 to Adams his plan of drawing the line to the Pacific along the
 Forty-first Parallel. Melish (in all but the first of the 1816 edi-
 SJefferson to Melsh, December 81, 1816, ia Jeferson papers, Division of
 Manuscripts, Library of Congres, voL 209.
 SMelish, Geographal Desertption (1818 ed.), pp. 20-45.