part iii.; but Raemdonck (p. 257) says he has never seen a copy of that

part which has it.

Mercator’s maps were followed, however, pretty closely in Mathias Quad’s or Quadus’s _Geographisch Handtbuch_,[732] Cologne, 1600, which contained a map of the world and another of North America, with some other special American maps; and such were also contained in the Latin version called _Fasciculus geographicus_, Cologne, 1608, etc.

In 1604 Mercator’s plates fell by purchase into the possession of Jodocus Hondius,[733] of Amsterdam, who got out a new edition in 1606,[734] to which he added fifty maps, including a few American ones; and thus began what is known as the _Hondius-Mercator Atlas_. The text was furnished by Montanus,[735] and the new maps were engraved by Petrus Kærius, who also prepared for Hondius the _Atlas minor Gerardi Mercatoris_ in 1607.[736]

After the death of Jodocus Hondius, Feb. 16, 1611, Heinrich Hondius (b. 1580; d. 1644) and Johannes Jannsonius (d. 1666) completed the _Atlas_; and what is known as the fourth edition (1613) contains portraits of Mercator and the elder Hondius. In this there were ten American maps, and for several editions subsequently there were 105 of Mercator’s maps and 51 of Hondius’. Such seemingly was the make-up of the seventh edition in 1619 (though called fourth on the title); but there is much arbitrary mingling of the maps observable in many copies of these early editions.

The same Latin text and its translations appeared in the several editions down to 1630, when what is called sometimes the eleventh edition appeared with 163 maps (105 by Mercator, 58 by Hondius); but I have noted copies with 184 maps, of which ten are American, and a copy dated 1632, with 178 maps. Raemdonck does not venture to enumerate all the Latin editions of Hondius and Jannsonius; but he mentions those of 1612, 1613, 1616, 1623, 1627, 1628, 1630, 1631.

In 1633 a marked change was made in the _Mercator-Hondius Atlas_. There was a new Latin text, and it was now called the _Atlas novus_, and made two volumes, containing 238 newly engraved maps (only 87 of Mercator’s remaining, while Hondius added 151, including 10 new maps of America). The French text was issued the same year, but it added details not in the Latin, and in the general description of America is quite different.[737] The German text also appeared in 1633; but it had—at least in the copy we have noted—only 160 maps, and of these 6 were American. The Dutch text is dated usually in 1634.

In 1635 the English text appeared with the following title: _Historia Mundi; or, Mercator’s Atlas.... Lately rectified in divers places, and also beautified and enlarged with new mappes and tables by the studious industry of Iudocus Hondy. Englished by W. S._, London;[738] and of this there was a second edition in 1637. The only map showing New France is a general one of America, which is no improvement upon that of the 1613 edition.

The English market was also supplied with another English version, published much more sumptuously, in two large folios, at Amsterdam in 1636, with the title, _Atlas; or, a Geographical Description of the Regions ... of the World, represented by New and Exact Maps. Translated by Henry Hexham. Printed at Amsterdam by Henry Hondius and John Johnson_.[739] The American maps are in the second volume, where the map of the two Americas is much like the world-map in vol. i. There is no part of New France shown in the special maps, except in that of “Nova Anglia, Novum Belgium, et Virginia,” where lying west of the Lac des Iroquois (Ontario) is a single and larger “Grand lac.”

A still further enlargement of the Mercator-Hondius _Atlas novus_ took place in 1638, when it appeared in three imperial folio volumes, with 318 maps, 17 of which are special maps of America.[740] It was now more commonly known as Jannson’s _Atlas_,—this publisher being a son-in-law of Jodocus Hondius,—and it went on increasing till it grew to eight volumes, to which were added a volume “Orbis Maritimus” (1657), a second on the ancient world, a celestial atlas for a third, and an “Atlas Contractus,” or _résumé_, for the fourth; making twelve in all.[741]

* * * * *

At this time there was a rival in the _Atlas_ of Blaeu, of which the reader will find an account in chapter ix. of the present volume, to be supplemented by the present brief statement.

Willem Jannson Blaeu was born in 1571, and died in 1638, and, with his sons Jean and Cornelis, devoted himself with untiring assiduity to his art. In 1647 the number of their maps reached one hundred. In 1655 their _Atlas_ had reached six volumes, and contained 372 maps. In this year (1655) the Blaeu establishment issued separately the American map, _Americæ nova Tabula_, with nine views of towns and representations of native costumes, accompanied by four pages of text. The Latin edition of 1662-63, _Atlas major, sive cosmographia Blaviana_, had 586 maps, of which the collection in the _Carter-Brown Catalogue_ (ii. 900) shows 23 in vol. xi. to belong to America.[742]

The Blaeu establishment was burned in 1672, and most of the plates were lost. Those which were saved passed into the hands of Frederic de Witt, who put his name on them, and they continued to be issued thus inscribed in the _Blaeu Atlas_ of 1685, etc.; and when De Witt’s business fell to Covens and Mortier, the inscriptions were again altered.[743]

* * * * *

A French atlas began a little later to attract attention, and ultimately made the name of its maker famous in cartographic annals. It was begun in 1646 by Nicolas Sanson d’Abbeville, who in 1647 was appointed Royal Geographer of France, and held that office till his death.[744] The volume of his _Atlas_, containing fifteen American maps, and entitled _L’Amérique, en plusieurs Cartes nouvelles et exactes_, was published by the author in Paris without date, but probably in 1656, though some copies are dated in 1657, 1658, and 1662.[745]

The elder Sanson, having been born in 1600, died in 1667, leaving about four hundred plates to his sons, who kept up the name,[746] and their stock subsequently fell to Robert Vaugondy, who has given a notice of the Sansons in his _Essai sur l’Hist. de la Géog._, as has Lenglet Dufresnoy in his _Méthode pour étudier la Géographie_.[747]

A new Dutch atlas, that of N. Visscher, called _Atlas minor, sive Geographia compendiosa_, appeared at Amsterdam about 1670. It contained twenty-six maps, and had three American maps; but the number was increased in later editions.[748] In 1680 it appeared in two volumes with 195 maps, 10 of which were American, and plates by Jannson, De Witt, and others, were included. It is not easy to discriminate among various composite atlases of this period, the chief cartographers being made to contribute to various imprints. Another _Atlas minor, novissimas Orbis Terrarum Tabulas complectens_, is likewise of this date (1680), and passes under the name of S. Wolfgang, with maps by Blaeu, Visscher, De Witt, and others. This usually contains nineteen American maps. Other atlases have the name of Frederic de Witt, who, as we have seen, got possession of some of Blaeu’s plates. The first example of his imprint appeared about 1675, at Amsterdam, with a printed index calling for 102 maps. Another edition (? 1680) is indexed for 160 plates, contained in two volumes of maps, and a third of charts.[749] Another small German atlas, the _Vorstellung der gantzen Welt_, of J. U. Muller, was published at Ulm in 1692, which had eighteen small American maps; and towards the close of the century the _Atlas minor_ of Allard obtained a good popularity. The pre-eminent name of Delisle, just becoming known, marked the opening of a new era in cartography, which is beyond the limits of the present volume.

* * * * *

Some notice should be given of another class of atlases, the successors of the portolanos of the sixteenth century, and the beginning of the later science of hydrography. In these the Dutch were conspicuous; and many of their subsequent charts trace back to the larger _pascaart_ of the North Atlantic which Jacob Aertz Colom published at Amsterdam about 1630.[750] Among the earliest of the regular _Zee-Atlases_ was that of Theunis Jacobsz, published in Amsterdam about 1635, which has a chart showing the American coast-line from Nova Francia to Virginia. Of large importance in this direction was the _Arcano del Mare_ of Robert Dudley, issued at Florence in 1646-1647, of which mention has been made in other chapters in this and in the preceding volume. Another of the Amsterdam Coloms—Arnold Colom—published his _Zee-Atlas_ about 1650, which contains six American coast-charts, and sometimes appears with a Latin title, _Ora maritima Orbis universi_, and is of interest in the historical study of our American coast-lines, improving as he does the preceding work of Jacobsz. Later editions of Colom, dating the charts, appeared in 1656 and 1663.[751] Of about this same date (1654) is a _pascaart_, published at Amsterdam, which seems to have been the joint business project of Frederic de Witt, Anthony and Theunis Jacobsz, and Gulielmus Blaeu. The world-map in it is dated 1652, and is doubly marked “C. J. Visscher” (Claes Jannson Visscher) and “Autore N. J. Piscator” (Nicolas Joanides), as the Latin equivalent of the same person. It shows the Atlantic coast from Labrador to Brazil. The first edition of Hendrick Doncker’s _Zee-Atlas ofte Water-Waereld_ appeared at Amsterdam in 1659, and is particularly useful for the American coasts. New maps were added to it in the edition of 1666; but the _Nieuwe Groote vermeerderde Zee-Atlas_ of 1676, though still called Doncker’s, is based on Colom, and has Colom’s six American charts. Additional American and other charts were added to the 1697 edition; while a set of still larger charts constitute Doncker’s _Nieuw Groot Zeekaert-boek_ of 1712.[752]

The _Zee-Atlas_ of Van Loon, with its forty-five double charts, appeared in 1661.[753] It is in parts reproduced from Blaeu, De Laet, and Jannson. Its numbers 46 and 47 show the coast from Newfoundland southwards. P. Goos, in his _Lichtende Colomme_, Amsterdam, 1657, had touched the Arctic coasts of America; but in his _Zee-Atlas_ of 1666 he gave in excellent manner eleven charts of the coasts of both Americas, out of the forty-one charts in all. These were all repeated in the edition of 1668-1669, and in the French edition, _Atlas de la Mer_, 1673. Other Dutch editions, with some changes, followed in 1675 and 1676. It was issued with an English text at Amsterdam in 1670.

Frederic de Witt, who had earlier appended to his _Atlas_ a section of maritime charts, published his _Zee-Atlas_ in 1675, which contained twenty-seven charts, eight of which were American; and in 1676 Arent Roggeveen issued his well-known navigator’s chart-book, which in English is known as _The Burning Fen_ (1676), and which also has a Spanish dress (1680). It gives in successive charts the whole eastern coast of the two Americas, on a large scale. Johann van Keulen, who had published a chart of the coast from Nantucket to Trinidad in 1680, issued a _Zee-Atlas_ in 1682-1687, based in part upon Van Loon, enlarging it in successive issues, so that in the edition of 1694 it had 146 charts, of which 38 were American. A later edition in 1734 contained 12 large folded charts of American coasts.[754]

Near the close of the century we come to the earliest of the French marine atlases, the _Neptune Français_, which Jaillot published in its enlarged form in 1693; but not till a _Suite du Neptune Français_ was issued in 1700 did any charts of American coasts make part of it. This contained eleven on America, professing to be based on Sanson’s drafts.

THE MAPS OF THE SEVENTEENTH CENTURY, SHOWING CANADA.

BY THE EDITOR.

[Detailed maps of the Upper Lakes and the Mississippi Basin, as well as those produced by Hennepin, though connected with this period, are made the subject of separate treatment elsewhere in the present volume. The general atlases are treated in the next preceding pages.]

IN the notes at the end of chapter ii. we followed the cartography of New France down to the opening of the seventeenth century. We saw in the map of Molineaux (1600) an indication of a great inland sea, as the prototype of the Great Lakes; but the general belief of the period, just as Champlain was entering on his discoveries, is well shown in the map, “Americæ sive Novi Orbis nova Descriptio,” which appeared in Botero’s _Relaciones universales_, published at Valladolid in 1603.[755]

The Spanish and the Dutch only repeated, but hardly with as much precision, what the map in Botero had shown;[756] and we only get approximate exactness when we come to the map of Lescarbot in 1609, of which sections are given in the present and in other chapters.[757] Champlain’s first map was made in 1612, and his second in 1613,[758] both of which appeared in _Les Voyages du Sieur de Champlain_, Paris, 1613. Between the issue of these 1612 and 1613 maps of Champlain and his greater one in 1632, the cartography of New France is illustrated by several conspicuous maps. Those of Hondius and Mercator, so called, of the same year were of course unaffected by the drafts of Champlain. We begin to notice some effects of Champlain’s work, however, in several of the Dutch maps; in that of Jacobsz, or Jacobsen, of 1621, for instance, of which account will be found on another page.[759] Maps by Jodocus Hondius and Blaeu represent a number of streams flowing from small lakes uniting to form the St. Lawrence. One by Jannson, in 1626, nearly resembles for the St. Lawrence region that portion of a “new and accurate map of the world, 1626,” which makes part of Speed’s _Prospect of the most famous Parts of the World_.

In 1625 the _Pilgrimes_[760] of Purchas introduces us to two significant maps. One is that which Sir William Alexander issued in his _Encouragement to Colonies_ in 1624, and was reproduced by Purchas, calling it “New England, New Scotland, and New France.” The essential part of it is given in Vol. III. chap. ix. The other is that called “The North Part of America,” ascribed to Master Briggs.

In the original edition of De Laet’s _Nieuwe Wereldt_,[761] published in 1625, we have a map of North America; but in the 1630 (Dutch) edition we find a special map of New France, which was repeated in the (Latin) 1633 edition. Harrisse[762] is in error in assigning the first appearance of this map to the 1640 French edition.

Champlain’s great map appeared in his 1632 edition.

It will be observed that Champlain had reached, in his plotting of the country east of the Penobscot, something more than tolerable accuracy. Farther west, proportions and relations were all wrong. The country between the St. Lawrence and the Gulf of Maine is much too narrow. The Penobscot is made almost to unite with the more northern river; and this error is perpetuated in the Dutch maps published by Blaeu, and Covens and Mortier many years later. The placing of Lake Champlain within a short distance of Casco Bay was another error that the later Dutch cartographers adopted in one form or another. Lake Ontario is not greatly misshapen; but Erie is stretched into a strait, while beyond a distorted Huron a “grand lac” is so placed as to leave a doubt if Superior or Michigan was intended.

Notwithstanding this pronounced belief in large inland seas, and the publication of the belief, the notion did not make converts in every direction. Two years later (1634) a map of Petrus Kærius, and even his other map, which appeared in Speed’s _Prospect of the most famous Parts of the World_, published in London, gave no intimation of Champlain’s results. The same backwardness of knowledge or apprehension is apparent in the map which accompanies the Amsterdam edition of Linschoten in 1644; in that of the world, dated 1651, which appeared in Speed’s 1676 edition; in the map in Petavius’s _History of the World_, London, 1659; and in two maps of N. I. Visscher, both dated 1652, which make the St. Lawrence River rise in the neighborhood of the Colorado. We might not expect the _Zee-Atlas_ of Van Loon to give signs of the inland lakes; but it is strange that the map “Americæ nova descriptio,” ignoring the great interior waters, was used in editions of Heylin’s _Cosmographie_, in London, from 1669 to 1677.

Some of the Dutch cartographers were not so inalert. Johannes Jannson in his _America septentrionalis_, and even Visscher himself in his _Novissima et accuratissima totius Americæ Descriptio_ give diverse interpretations to this idea of the inland seas. The draft in the Hexham English translation (1636) of the Mercator-Hondius atlas is not much nearer that of Champlain.

Harrisse (_Notes_, etc., nos. 190, 191) refers to two charts of the St. Lawrence of 1641 which are preserved in Paris, and are known to be the work of Jean Bourdon, who came to Quebec in 1633-34. Perhaps one of these is the same referred to by Kohl, as dated 1635, and in the _Dépôt de la Marine_, of which a copy is in the Kohl Collection in the State Department at Washington. Harrisse also (no. 324) refers to a _Description de la Nouvelle France_,—a map published by Boisseau in Paris in 1643.

The map in Dudley’s _Arcano del Mare_ (Florence, 1647), called “Carta particolare della terra nuova, con la gran Baia et il Fiume grande della Canida: D’America, carta prima,”[763] presents a surprise in making the St. Croix River connect the Bay of Fundy with the St. Lawrence; and Dudley seems to have had very confused notions of the sites of Hochelaga and the Saguenay. The annexed sketch is much reduced.

The same transverse strait appears in _Carte générale des Costes de l’Amérique_, published at Amsterdam by Covens and Mortier. A treatment of the geographical problem of the lakes which had more or less vogue, is shown in Gottfried’s _Neue Welt_, 1655, in a map called “America noviter delineate;” and this same treatment was preserved by Blaeu so late as 1685.

A most decided advance came with the map, _Le Canada, ou Nouvelle France_, of Nicolas Sanson in 1656,[764]—a far better correlation of the three lower lakes than we had found in Champlain, with an indication of those farther west.[765] Contemporary with Sanson was the English geographer Peter Heylin, whose map, as has already been noted, betrays no knowledge of Champlain. His _Cosmographie in Four Books_ appeared in 1657,[766] and the second part of the fourth book relates to America, and is accompanied by the map in question. The contemporary Dutch maps of Jannson, Visscher, and Blaeu deserve little notice as contributions to knowledge.[767]

Of the map of Creuxius, made in 1660 and published in 1664, a fac-simile of a part is annexed.[768] For the eastern parts of the country reference may be made to the map _Tabula Novæ Franciæ_, of about 1663, given in the chapter on Acadie.[769]

One of the volumes of the great _Blaeu Atlas_ of 1662, _America, quæ est Geographiæ Blavianæ Pars quinta_, very singularly ignored all that the cartographers of New France had been long divulging, and the same misrepresentation was persistently employed in the later _Blaeu Atlas_ of 1685, which contained in other American maps a variety of notions equally erroneous, and which had been current at a period very long passed.

The map in Montanus’s _De Nieuwe en Onbekende Weereld_, 1670, “per Jacobum Meursium,” not the same as the “Novissima et accuratissima totius Americæ Descriptio” of John Ogilby’s great folio on _America_, 1670, and later years, seems to be substantially N. Visscher’s map of the same title, issued in Amsterdam in the same year.[770]

The maps of Hennepin (1683-1697) form a part of a special note elsewhere in the present volume; and the map accompanying Le Clercq’s _Etablissement de la Foy_, 1691, is also reproduced in Shea’s translation of that book.[771] It makes the Mississippi debouch on the Texas shore of the Gulf of Mexico, as many of the maps of this period do.

Maps of a general character, indicating a knowledge of the interior topography of America, sometimes expanding, and not seldom retrograde, followed rapidly as the century was closing, of which the most important were the maps of _Amérique septentrionale_ (1667, 1669, 1674, 1685, 1690, 1692, 1695), by the Sansons, and the Roman reprint of it in 1677,[772] as well as _La Mer du Nort_ of Du Val in 1679,[773] Sanson’s _Le Nouveau Mexique_, of the same year, which extends from Montreal to the Gulf;[774] the _North America_ of the English geographer, William Berry (1680);[775] the _Partie de la Nouvelle France_ of Hubert Jaillott (1685);[776] and the same cartographer’s _Amérique septentrionale_ of 1694, and _Le Monde_ of 1696; the _Carte Generalle de la Nouvelle France_[777] (1692) engraved by Boudan; the _Amérique septentrionale_ of De Fer (1693); the marine _Cartes_ (1696) of Le Cordier;[778] the _New Sett of Maps_ published by Edward Wells in London in 1698-99; and finally the _Amérique septentrionale_ of Delisle.[779] The maps of La Hontan (1703-1709) are the subject of special treatment in another note.

If we run through the series of maps here sketched, we cannot but be struck with the unsettled notions regarding the geography of the St. Lawrence Valley. Beginning with the clear intimation by Molineaux, in 1600, of a great body of interior water, which was the mysterious link between the Atlantic and the Arctic seas, and finding this idea modified by Botero and others, we see Champlain in 1613 still leaving it vague. The maps of the next few years paid little attention to any features farther west than the limit of tide-water; and not till we reach the great map which accompanied the final edition of Champlain’s collected voyages in 1632 do we begin to get a distorted plot of the upper lakes, Lake Erie being nothing more than a channel of varying width connecting them with Lake Huron. The first really serviceable delineation of the great lakes were the maps of Sanson and Du Creux, or Creuxius, in 1656 and 1660. Here we find Lake Erie given its due prominence; Huron is unduly large, but in its right position; and Michigan and Superior, though not completed, are placed with approximate accuracy. This truth of position, however, was disregarded by many a later geographer, till we reach a type of map, about the end of the century, which is exemplified in that given by Campanius in 1702.

A water-way which made an island of greater or less extent of the peninsula which lies between the St. Lawrence and the Atlantic, appeared first in 1600 on the Molineaux map, and was repeated by Dudley in 1647; but on other maps the water-sheds were separated by a narrow tract. So much uncertainty attended this feature that the short portage of the prevailing notion was far from constant in its position, and on some maps seems repeated in more than one place,—taking now the appearance of a connection on the line of the St. Croix, or some other river of New Brunswick; now on that of the Kennebec and Chaudière; again as if having some connection with Lake Champlain, when a misconception of its true position placed that expanse of water between the Connecticut and the Saco; and once more on the line of the Hudson and Lake George.