290 DEVELOPMENT OF THE CARTOGRAPHY OF AMERICA. 



first, under the equator. The eastern projection of Brazil, Cape St. 

 Augustine, appears early in its proper situation. Far more difficult 

 than the latitude was the determination of the longitude. They could 

 not then reckon astronomically. It was much easier to ascertain the 

 dangerous character of the coast through an estimate of the rapidity 

 of the passage. But that even here opinions differed may be per- 

 ceived from the varying estimates of pilots on the occasion of the first 

 voyage of Colon. His own observations or calculations of latitude are 

 no better than those of hi& contemporaries. Nor do 1 believe that the 

 few astronomical efforts that were made to determine the longitude on 

 American soil have had any inlluence upon the maps. It is therefore 

 not to be wondered at that the eastern shores of the New World, 

 namely, South and Central America, should be removed 3 to 5 degrees 

 too far toward the east. Things were even in a worse state in the 

 beginning in the case of North America. The insular-like appearance 

 of the shores of Newfoundland, Labrador, and Greenland (?) scarcely 

 permit us, at first, to devine their actnal character. The eastern coast 

 of the present United States was quite delusive, extending horizontally 

 from west to east, as if the ocean were quite dammed off on the north. 

 As a necessary sequence, Newfoundland lay about 40 degrees too far 

 to the east. These errors had undergone hardly any correction during 

 the sixteenth century. 



The line of demarcation, decided upon since 1494, frequently served 

 on the general charts as an initial meridian. We first meet with it on 

 Sheet IV, in Kuustmann's Atlas, and consequently somewhere about 

 the year 1518. It here lies 21 to 22 degrees westward from the Cape 

 Verde Island Santo Antonio, and hence 370 leagues remote from it, as 

 was settled in the convention of Tordesillas in 1494. liut Spanish 

 cosmographers wished to remove it as far as possible to the east, in 

 order to lessen the domains of Portugal. They were unable to come 

 to a full settlement of this point. The old charts, whether of land or 

 sea, track charts, or maps multiplied by typography, could alone supply 

 a scientific desideratum where the date of restoration is clearly indi- 

 cated by the compiler, and when, moreov^er, an exact computation of 

 the time is made, or where this date may be ascertained with reasonable 

 probability from the text. The greater portion of the old cartographic 

 records that have been x)reserved are without date. 



Of the twelve oldest sheets that we possess from the year 1500 to 

 1509, four are dated and eight are without date. We may regard it as 

 a great piece of good fortune that the oldest chart of Juan de la Cosa 

 that has been preserved bears a precise date, and that to the so-called 

 Cantino chart a definite year may be ascribed with confidence through 

 the letter that accompanies it. In other cases a determination of the 

 time of a chart without date is always an extremely difficult problem. 



It is true that in many instances it is possible to verify Avliich are the 

 most recent discoveries recorded, but it follows necessarily from this 



