The working-method of the new cartographers: 

The Gulf of Mexico and Spanish sources, 1696-1718

Monique Pelletier

      The lands bordering the Gulf of Mexico attracted the attention of Claude (1644-1720) and Guillaume (1675-1726) Delisle around 1700, after the voyages of Robert Cavelier de La Salle and at the time when the French began to establish themselves in the lower Mississippi valley. This had international consequences, for the Spanish king Philip V protested, declaring that the Mississippi River was the greatest jewel in his crown.  In order to work out the cartography of the region, the Delisles generally relied on Spanish textual sources, since the cartographic evidence was both lacking in interest and also hard to obtain.  The printed work of Guillaume is well known, but the more abundant manuscript work of the Delisles, father and son, has been less studied, even if it has been listed and cited by researchers like Jack Jackson and Nelson-Martin Dawson.

      The cartographic collection belonging to the “Fonds Delisle” at the Archives nationales de France in Paris contains both geographical sketches, relating to texts coming from the Delisles’ reading, and also maps, drawn up by Guillaume.  For the lands surrounding the Gulf of Mexico, the maps show both the development of the cartographer’s ideas and his hesitancy, while the sketches rely on the content of those sixteenth- and seventeenth-century Spanish works translated into French:  the travel narrative of Hernando de Soto for Florida (the future Louisiana), and the works of Alonso Benavides for New Mexico and of Antonio de Herrera for New Spain.  We can trace the delicate mixture of textual and cartographic sources by seeing how the sketches  were incorporated into the maps, with the additional use of a Spanish portolan chart.  This process cannot be seen apart from the aims and development of the colonial policy of the French monarchy, which also supplied information to the cartographers.

      Guillaume Delisle published maps of America from 1700 onwards.  At that time he had not yet been elected to the “Académie royale des Sciences,” but he would be in 1702 as a student of astronomy, “even though he was not an observer.”  In fact, he had received a joint training: one type with his father Claude, the historian/geographer for whom he prepared maps and globes, and another with Jean-Dominique Cassini, who taught him astronomy.  It was necessarily the latter who persuaded him to use the figures of latitude and longitude set out by the Académie des Sciences, to compile documents which would renew French cartography.

      A recent work¹ refers to the Delisles as continuing the work of Nicolas Sanson, which is true as far as their pedagogical effort is concerned, but which is more debatable as far as method goes.  For while Sanson used existing cartographic models, which he developed further using textual information, the Delisles began by considering the validity of the models, renewing them by incorporating new observations. When such observations were lacking, they did not neglect any other source of information, whether old or new, but tried to work out new cartographic models using available texts.  Nicolas Sanson had already used textual information, following a system that he may have passed on to Claude Delisle,² and that may emerge from the cartographic sketches that we are going to consider.  The examples shown here will demonstrate the method followed by the Delisles, whose printed maps give only a faint idea of their intense activity in re-working their cartographic ideas.  They were very different from the Sanson dynasty, who produced a great number of varied printed maps, but did not often review them.

From Florida to Louisiana

      The Delisles, like other French cartographers, wondered where they should plot the mouth of the Mississippi River, reached in 1682 by Robert Cavelier de La Salle.  He had come right down the great river, and in the name of the king of France had  taken possession of a “Louisiana” stretching from the Ohio River to the Gulf of Mexico and, to the west, as far as the mouth of the “Rivière des Palmes”, that is to say as far as the frontier of New Spain.  This is what is shown on the map by Roussel which accompanied the Description de la Louisiane published in 1683 by Father Louis Hennepin, who claimed himself to have discovered the upper valley of the Mississippi River (fig. 1).

      The region called “Louisiana” by La Salle would interest Louis XIV if it were near the Spanish mines of New Mexico, as the explorer seems to think.  This proximity links up with the ideas of abbé Claude Bernou who, even before La Salle’s voyage, had in January 1682 sent to the marquis de Seignelay (Colbert’s son), a plan for colonization from the mouth of the Rio Bravo—on the coast of what he still calls “Florida”—towards the silver, gold and lead mines of Nova Biscaya.³  These toponyms—Florida, Nova Biscaya and Rio Bravo—appear on the splendid map attributed to Bernou, who restricted the name “Louisiana” to that part of the Mississippi River explored before 1682.⁴  La Salle’s second voyage, of 1684, ended tragically in 1687 without having found the mouth of the great river.  French exploration of Louisiana resumed after the peace of Ryswick in 1697, under the patronage of the naval minister, Louis de Pontchartrain.  He appointed Pierre Le Moyne d’Iberville head of a new expedition which, working along the coast, detected the mouth of the Mississippi River on 3 March 1699.⁵

      This was the situation when Claude Delisle published in Le Journal des Savans for 17 May 1700 a letter to M. de Cassini concerning the mouth of the Mississippi (“Lettre […] à M. de Cassini sur l’embouchure de la riviere de Mississipi”).  This letter is well known, and sets out the near impossibility of locating the position of the mouth with precision, because astronomical methods – observing the eclipses of the satellites of Jupiter or of the moon – could not have been used and because the coast of Florida is one “of the least known in America” (“des moins conües de l’Amerique”).  So Claude Delisle compiled his map of the coast and interior of Florida (“carte de l’intérieur et de le côte de Floride”) using well-known travel accounts which had been both published and often translated into French.  Thus he quoted the expeditions of Pánfilo de Narváez (c. 1460-1529), of Alvar Nuñez Cabeza de Vaca and of Hernando de Soto (1500-1542), emphasizing the difficulty of using textual sources for cartographic work.

      In fact, some of these sources are difficult to use, as the travel accounts of Hernando de Soto show.⁶  The one compiled by Garcilaso de la Vega (1539-1616) is more literary than precise, and it is the one written by a Portuguese gentleman from Elvas, published at Evora (Portugal) in 1557 and translated into French in 1685,⁷ that the Delisles seem to prefer.⁸  They made a summary of it, and drew sketches carefully following the Portuguese gentleman’s instructions⁹ (fig. 2), the result of which is far from satisfactory.

      Also, in order to work out the various stops made by de Soto that are shown on the manuscript 1696 map of New France and neighboring countries¹⁰ (fig. 3), Guillaume combined the Portuguese gentleman’s toponymy with information on directions and distances from the account of Garcilaso de la Vega.   One of the main points at issue was the exact position of Ucita, from which the expedition had set out.  The first sketches put it on a bay of the northern coast of the Gulf of Mexico, between the Florida peninsula and the mouth of the Mississippi.  The manuscript map of 1696 has two bays “of Saint-Esprit”; one of them, called the “Baye du St Esprit ou Tacobago,” is on the western coast of the Florida peninsula and was thought to be de Soto’s base.  Another manuscript map, which in its coastal outlines resembles Guillaume  Delisle’s Amerique Septentrionale of 1700, suggests a more northerly point of departure, to the northwest of the same peninsula (fig. 4).

      Finally, the cartographer came back in 1718 to his interpretation of 1696.  He also had to relate the explorer’s itinerary to the hydrography of the region, and particularly to the Mississippi River.  On the 1696 map, then, Delisle could put in little sections of the rivers crossed by de Soto, but he clearly could not draw the rivers fully.  The manuscript map drawn about 1700 was more carefully prepared, perhaps for Jérôme de Phélypeaux, comte de Pontchartrain, naval secretary of state between 1699 and 1715.  On it, Delisle retained the “mer de l’ouest”, kept the name “Loüisiane” alongside the Mississippi River—a name which would disappear from the map of 1700—and made it exist together with that of “Floride,” above the Gulf coast.

      The itinerary of Hernando de Soto may also be seen on the Carte de la Louisiane et du cours du Mississippi, published by Delisle in 1718.  It is the map of a greater Louisiane containing the most recent travels, those of Cavelier de La Salle (1687), of Henri de Tonty (1702), of Louis Juchereau de Saint-Denis (1713 and 1716), and of Jean-Baptiste Le Moyne de Bienville.  As on the map of 1696, de Soto is shown as leaving from the west coast of the Florida peninsula (at a latitude about that of Cape Canaveral) and following a track—close to that suggested by Delisle’s first version—whose stops are marked by stars.  From there on, the explorer’s route is inserted on a more complete map that included the course of several rivers.  Thus, it no longer serves to construct the map, but rather to give the document a historical foundation.

Figure 5. Florida and New Mexico on the Carte due Mexique et de la Floride published by Guillaume Delisle in 1703.

Figure 6. Eastern part of the great manuscript map of the Mississippi area compiled by Guillaume Delisle about 1703 (Archives Nationales, Paris).

      Guillaume Delisle emphasized the importance of the map published in 1703¹¹ under the title Carte du Mexique et de la Floride, des terres angloises et des isles Antilles, du cours et des environs de la riviere de Mississippi (fig. 5).  He claimed that a larger-scale version of this document had been presented at the court, and this is perhaps the Carte des environs du Mississippi preserved at the Archives nationales¹² (fig.6).  The de Soto travel accounts still appear among the sources for the 1703 map, but Delisle does not put in the itinerary.  He suppresses the name “Louisiane,” stretching “Floride” out between Canada (or “Nouvelle France”) to the north, the English colonies to the east, and “Nouveau Mexique” to the west.  The published map shows the Spanish establishments on the coast to the east of the Mississippi:  Pensacola, Saint-Joseph and Sainte-Marie d’Apalache, while only “Apalachicoly” and the “Apalaches” appear on the large manuscript map.  This also contains soundings found as well on Nicolas de Fer’s Partie meridionale de la riviere de Missisipi, published in 1718 on the basis of Delisle’s map of 1703.¹³  None of these maps shows the fort of Mobile, whose construction had all the same been hurried forward by Le Moyne d’Iberville in 1701.  As he shows only the fort of Biloxi (1699) and that of Missisippi (1700), Delisle must have relied on the account of d’Iberville’s second voyage during which the explorer made soundings at the entrance to the Spanish port of Pensacola and in Mobile Bay.  It was also during this voyage that Charles Le Sueur delineated the course of the Mississippi River, a version that was used by Delisle for compiling his Carte de la riviere de Missisipi of 1702,¹⁴ which has remained manuscript, and for inserting the river on the great manuscript map now conserved at the Archives nationales.

      The position of official cartographers like the Delisles allowed them access to confidential documents.  Thus Guillaume Delisle claimed that for his 1703 map he used a portolan chart that he does not identify, but that I believe I have found in the “Service hydrographique” documents now held in the department of maps and plans at the Bibliothèque nationale de France.¹⁵  If I am correct, the author of this document was the Spanish pilot Juan Bisente del Campo, and it was compiled at Havana and dedicated to admiral Guillermo Morphi in 1696.¹⁶   Bisente had drawn another chart covering the same area at Cartagena in November 1700, and this document also ended up in France.¹⁷  The 1696 chart may have been captured from the Spaniards in 1697 by the royal ship Le Bon in the West Indies, and if this is so, then it would have fallen into the hands of Mr. de Patoulet, the ship’s captain.  This at all events is what a note on another manuscript map – perhaps a copy of the first chart – suggests.¹⁸  Another copy¹⁹ (fig. 7), more similar to this portolan to which it explicitly refers, may, in the absence of the original, have been used by Guillaume Delisle to review the shape of the western coast of the Gulf from the mouth of the Mississippi River to Yucatan.  This revision could only have been made using either the work of Bisente or some very similar chart; moreover, either the original or the copy of the 1696 chart must have helped Lemoyne d’Iberville to find the mouth of the great river in 1699.

“Le Nouveau Mexique”

      This area appeared on the 1703 map alongside Florida.  In order to show it on the previous maps, Guillaume Delisle had already used the account of Alonso Benavides, published at Madrid in 1630 and translated into French in an edition that came out at Bruxelles in 1631 under the title of Requeste remonsrative [sic] au roi d’Espagne sur la conversion du Nouveau Mexico.  This text was also summarized in the collection by Johannes de Laet published in French at Leyden in 1640.  It allowed Delisle to plot the position of the Indian tribes of Nouveau Mexique in the positions suggested by Benavides, and to estimate certain distances.  On a sketch (fig. 8) the Delisles set out the geographical position of information derived from the text of the Spanish Cordelier.²⁰  There is, for instance, the following passage from de Laet:

Nouvelle Mexique, or at any rate its capital city, Santa Fé, is situated on the 27^th degree of the line to the north [Santa Fé is a little above 35 degrees on the Carte du Mexique of 1703 and at about 37 degrees on the Carte de la Louisiane of 1718]; you go there from the silver mines of S. Barbara through the province of the Conchos, which is separated from New Spain by a river of the same name.  Between S. Barbara and the “riviere del Norte” is about one hundred leagues, full of danger through the lands of the Tabosos, Tarrahumares, Tepoanes, Tomites, Sumas, Hanos and other cruel and savage peoples […].  From the riviere del Norte it is also one hundred leagues to Nouvelle Mexique,²¹ and here you first meet the Mansos and Gorretas […].  You come along the riviere del Norte, where Nouvelle Mexique begins and has one hundred leagues from Saint Anthonio de Senecu, the first township of the Biroros, as far as the township of S. Hieronimo in the province of the Taoros.²²

      The Delisles used this information first on the manuscript maps²³ that preceded the engraved map of 1700.  Thus the three manuscript maps and the engraved map show a zone in the upper valley of the rio del Norte, called also “Nouveau Mexique”, which illustrates the ideas of Benavides.  Based on Santa Fé, created by the Spaniards at the beginning of the seventeenth century, this region on Guillaume Delisle’s maps includes as well information derived from another map owned by the Delisles,²⁴ also used by Vincenzo Coronelli for Louis XIV’s great terrestrial globe.²⁵  This map was called the Carte du Nouveau Mexique tirée des relations de Mr le comte de Peñalossa qui en a été gouverneur en 1665, du manuscrit du P. Estevan de Peroa, custode de l’ordre de saint François dans le même pays et d’autres memoires écrits sur les lieux, and came from documentation assembled by abbé Claude Bernou.

      Bernou, related to Jean Bobé, another acquaintance of the Delisles,²⁶ is among the people proposed as expert witnesses in the case for cartographic plagiarism brought against Jean-Baptiste Nolin.²⁷  The comte de Peñalosa had come to the attention of the abbé Bernou, who at one time wanted him to lead a colony proposed for New Biscaya.  After being chased out of Nouveau Mexique by the Inquisition, Peñalosa had successively offered his services to England and to France, proposing to conquer the north of the Spanish province, consisting of the lands of Quivira and Thogayo, which appear on Delisle’s map of 1703.  As was the case for Coronelli, the Delisles seemed to attract  adventurers who craved new space, and thought that cartography could give their projects a certain appeal and make them attractive to the king and his ministers.

“La Nouvelle Espagne”

      The Spanish chart signed by Bisente was thus used by Guillaume Delisle for delineating the coasts of Nouveau Mexique and of Nouvelle Espagne towards 1703.  But for the latter great province Delisle could also call on a precious text:  the Description des Indes occidentales of Antonio de Herrera.  He had been named historian for the Indies by Philip II, and after being published at  Madrid in 1601-1605, his work came out at Amsterdam in a French translation in 1622.  Relying on this text, the Delisles compiled a map called L’audience de la Nouvelle Gallice et de la Nouvelle Espagne,²⁸ and this was based on a set of distances which allowed them to form constructional triangles (fig. 9).  Herrera provided many figures of distances, as the following extract, used by Delisle in his preparatory map, shows:

The bishopric of Guaxaca, which takes its name from the province, also called Antequera, from the city with the cathedral, between the bishopric of Los Angeles and the bishoprics of Guatemala, contains 125 leagues from one sea to the other on the side of the diocese of Tlascala, and 60 towards that of Chiapa, and 100 along the coast of the southern sea, and 50 to the coast of the northern sea, including the provinces of upper and lower Mixteca […].  When the treasurer Alfonse d’Estrada was governor he settled the town of Saint Alfonse de los Zapotecas, 20 leagues from Antequera towards the north-east.²⁹

      One may also wonder if the Spaniard did not have one or more maps, much  more detailed than those which accompany his Description des Indes occidentales, and which were probably based on the sixteenth-century maps drawn by Juan López de Velasco.³⁰  Delisle unhesitatingly used the information from Herrera in his map of 1703, in addition to the coastal outline derived from Bisente’s chart.

Conclusion

     Our examination of three case-studies of Florida, Nouveau Mexique and Nouvelle Espagne, coming from the Delisle archives, and analysis of the corresponding maps drawn by Guillaume Delisle, show that the documentation was always consulted, but that it was differently used for different maps.  Thus the material concerning the de Soto voyage eventually lost its primary value and became only of historic interest.  On the other hand, the information generated from Herrera became crucially important when it was combined with the secret chart captured from the Spaniards.  As for information from Benavides and Peñalosa, it was used to locate Indian tribes and was checked for internal consistency as this note shows:  “the comte Peñalosa told the abbé Bernou that it was a mistake to put Quivira in the west of Nouveau Mexique and that it was surely in the east [the solution adopted by the 1703 map].  Benavides puts this city in  both east and west; Laet says that he does not have much faith in this Cordelier.”³¹

     The evolution of cartography cannot be studied apart from the history of the colonial policy directed by the royal power, which allowed the geographer to collect new information and so to create new documents for the monarch’s use.  What would have become of Guillaume Delisle’s American work, without the scientific support of Le Moyne d’Iberville and of Le Sueur, whose expeditions were a response  to the aims expressed by Louis XIV in 1697?  For the king, the Mississippi is the “one place where you can export the goods of Louisiana, which His Majesty’s  agents discovered some years ago, and which would be useless to him unless he were master of this river-mouth.”³²

     The objective of d’Iberville’s second voyage was to investigate  the advantages of establishing a colony on the banks of the great river.  However, questions about the successor of Charles II—who would die in 1700—made France prudent and discouraged conflict with the Spaniards over Pensacola.  Moreover, Louis XIV was interested in trading with the Spanish colonies in America.  In September 1698 he created the Compagnie de Saint-Domingue, one of whose objects was to organize interlopers in the trade with Terra Firma.  Once Louis XIV’s grandson had ascended the Spanish throne in 1700 with the title of Philip V, the former governor of Saint-Domingue, Jean Ducasse, negotiated with the Madrid government the treaty known as the Asiento, concerning the importation of slaves into French and Spanish America.  Thereafter he collaborated with Spain by ensuring the security of its galleons.  Then in 1701 Louis XIV made the colony on the lower Mississippi independent of New France.  All these events and objectives may be traced on the Carte du Mexique et de la Floride of 1703, which traces the frontiers of Florida, defines the Spanish possessions and does not forget to show the division of Saint-Domingue between France and Spain.  This is the island from which French expeditions into the Gulf left.

     Louisiana replaced “la Floride” of the 1703 map as a text of 1712 shows, as the term for a “country at present known as the government of Louisiana,” but now taking in the Ohio River.  Letters-patent empowered the sieur “Crozat to direct the trade in all our lands between Nouveau Mexique and the English in the Carolinas […] principally the Saint Louis river, once called Mississippi, from the mouth up to the Illinois [though not including them] […] with all the lands […] and the rivers which flow directly or indirectly into this part of the Saint-Louis river.”³³  But Crozat had been too ambitious and on 6 September 1717 further letters were given to the Compagnie d’Occident, allotting to it Louisiana, to which Louis XV soon added “the country of the Illinois.”³⁴  This was the context for Guillaume Delisle’s 1718 Carte de la Louisiane et du cours du Mississippi, published at the same time as the similar map of Nicolas de Fer, geographer to the king of Spain.  Like Delisle, de Fer took advantage of the political situation which encouraged the sale of his map, though it was compiled from old work.

     1718 was an important year in Guillaume Delisle’s career.  On 24 August he received the title of premier géographe du roi from the king to whom he and his father had given history and geography lessons.  As the citation put it, he had given “authentic proof” of his “deep erudition” and had published many geographical works of high quality.  Royal favor would allow him to continue publishing “such useful works.”³⁵  Our examination of a few examples concerning the Gulf of Mexico suggests that this was a wise appointment, honoring the father through the son, both having brought to their joint work on cartography the rigor of historical analysis.

¹   N.-M. Dawson, L’atelier Delisle: l’Amérique du Nord sur la table à dessin (Sillery, Québec: Éd. Du Septentrion, 2000).

²   Ibid., p. 24.

³   Bibl. nat. de France, Manuscrits, Clairambault 1016, fo. 206-7, Découverte et établissements des Français dans l’ouest et le sud de l’Amérique septentrionale, 1614-1698, ed. P. Margry, t. III (Paris, 1879), p. 44-48.

⁴   Bibl. nat. de France, Cartes et Plans, SH, pf. 122, div. 2, p. 0.

⁵   On the use of information from d’Iberville, see N.-M. Dawson, op. cit.

⁶   B. Boston, “The route of De Soto: Delisle’s interpretation,” Mid-America 21 (1939): 277-297.

⁷   Histoire de la conqueste de la Floride par les Espagnols sous Ferdinand de Soto, written in Portuguese by a gentleman of the town of Elvas and translated by S. de Broë, seigneur de Citry et de La Guette. Paris, 1685.

⁸   Extrait de la relation de la conqueste de la Floride,  Arch. nat., 2 JJ 55, IX, 3.

⁹   The sketches drawn according to the account of the Portuguese gentleman are conserved in 6 JJ 75, p. 231 C and B, where they are mistakenly described as coming from the account of Garcilaso de la Vega.

¹⁰   Arch. nat., 6 JJ 75, p. 130.

¹¹   See J. Delanglez, “The sources of the Delisle map of America, 1703”, Mid-America 26/4 (1944): 255-297.

¹²   Arch. nat., 6 JJ 75, p. 253 (manuscript document). See Jack Jackson, Manuscript maps concerning the Gulf Coast, Texas and the Southwest (1519-1836) (Chicago: The Newberry Library, 1995), p. 52.

¹³   In 1701, de Fer, who was then géographe for the dauphin, published a map of the Costes aux environs de la riviere de Misisipi,  in which he tried to illustrate the expeditions of La Salle and Iberville (1683-1699). This map gave a better location for the mouth of the Mississippi than that of the first versions of Delisle’s 1700 map.  It rather resembles information from the Bisente portolan chart, used by Delisle for the two maps that we have been discussing. On this portolan, see below, p. 9. All this makes one believe that Delisle’s work was ready before 1703.

¹⁴   Bibl. nat. de France, Cartes et Plans, SH, pf. 138 bis, div. 2, p. 3 (five leaves).

¹⁵   M. Pelletier, “Louis XIV et l’Amérique;  témoignages de la cartographie,” Bulletin du Comité français de cartographie 115 (March 1988): 58. J. Jackson, Manuscript Maps, p. 52.

¹⁶   Bibl. nat. de France, Cartes et Plans, SH, Archives 27.

¹⁷   Ibid., Archives 28.

¹⁸   Bibl. nat. de France, Cartes et Plans, SH, pf. 141, div. 1, p. 3, a copy which lacks coastal names and has notes on the La Salle expedition.

¹⁹   Ibid., p. 2.

²⁰   Arch. nat., 6 JJ 75, p. 276.

²¹   In the limited sense of the term, including the Spanish settlements around Santa Fé, and from S. Antonio de Senecu to S. Hieronimo, as shown on the map of 1696 and some later ones, even though the Carte de la Louisiane of 1718 did not keep these two place-names.

²²   “New Mexico according to the account of Alfonso de Benavides, Cordelier,” in J. Laet, Histoire du Nouveau Monde ou Description des Indes occidentales (Leiden, 1640), p. 233.

²³   Arch. nat., 6 JJ 75, p. 130, 128 (1 and 2).

²⁴   Arch. nat., 6 JJ 75, p. 270. This document was used when Nolin, in the 1680s, published the Coronelli map entitled Le Nouveau Mexique, Nouvelle Grenade, Marata, avec partie de la Californie.

²⁵   M. Pelletier, “Les globes de Louis XIV, les sources françaises de l’oeuvre de Coronelli,” Imago Mundi, 34 (1982): 83.

²⁶   N.-M. Dawson, L’atelier Delisle, p. 1127; Bobé is named the “abbé Bernov de la Régence.”

²⁷   N.-M. Dawson, op. cit., p. 35.

²⁸   Arch. nat., 6 JJ 75, p. 278 : L’audience de la Nouvelle Gallice et de la Nouvelle Espagne, from Herrera.

²⁹   A. de Herrera, Description des Indes occidentales (Amsterdam: Michel Colin, 1622), pp. 26-27.

³⁰   D. Buisseret, “Mexiko” in Lexikon zur Geschichte der Kartographie (Vienna: F. Deuticke, 1986), p. 491.

³¹   Quoted by N.-M. Dawson, L’atelier Delisle, p. 153.

³²   Quoted by M. de Villiers, “La Louisiane, histoire de son nom et de ses frontières successives (1681-1819),” Journal de la Société des Américanistes, New Series, XXI (1929): 40.

³³   Quoted by Villiers, op. cit.,  p. 44-45.

³⁴   Ibid., p. 46.

³⁵   From the title quoted by N.-M. Dawson, L’atelier Delisle, p. 44.