In Northern Mists Volume Ii Part 18

The credit of having introduced the name of Greenland, with the ancient Nors.e.m.e.n's geographical ideas about the extreme North, into cartography belongs, so far as is known, to the Dane Claudius Clausson Swart, usually called in Latin Claudius Clavus (sometimes also Nicolaus Niger). He was born in Funen, travelled about Europe, and, as shown by Storm [1891, pp.

17, f.], was probably the "Nicolaus Gothus" who is mentioned at Rome in January 1424, and who is reported to have there given out that he had seen a copy of Livy in the monastery of Soro, near Roskilde (which was probably a romance on his part). We are told that he was a man of acute intelligence, but a rover and unsteady. His subsequent history is unknown.

As a supplement to Ptolemy's Geography, which just at that time (1409) was becoming known in Western Europe in a Latin translation, he made, probably in Italy, two maps of the North, with accompanying descriptions. The maps must have been drawn either by himself or with his help. They are the first maps known in Western Europe which are furnished, after the model of Ptolemy (or Marinus), with lines of lat.i.tude and longitude,[243] and they thus mark the beginning of a more scientific cartography and geography in Western Europe.[244]

His first map (the Nancy map) must have been drawn between the years 1413 and 1427, probably between 1424 and 1427; but it can never have been widely known, as it has exercised no noticeable influence on the cartography of the succeeding period. The French cardinal Filastre (ob.

1428), who was staying in Rome in 1427, became acquainted with it there, and made a reduced copy of it, which, together with a copy of the accompanying text, he had bound up with his copy of the Latin translation of Ptolemy's Geography with maps. This work was not rediscovered at Nancy until 1835, when it was published; the map is therefore usually called the Nancy map. Clavus's second map, which seems to have been drawn later than that just mentioned, has on the other hand had considerable influence on the cartographical representation of the northern regions through a period of two centuries.

A copy of the later map was first brought to light by Nordenskiold at Warsaw in 1889 [1889, p. x.x.x.]; since then several copies have been rescued from oblivion, while the text accompanying the map was accidentally discovered in 1900 by Dr. A. A. Bjornbo in a mediaeval MS. at Vienna [Bjornbo and Petersen, 1904]. The original map is lost; but except as regards details of no great consequence there can now be no doubt as to what it was like.

The reproductions (pp. 248 and 251) will give an idea of the representation of the North on the two maps. As far as Ptolemy's map extended (cf. vol. i. pp. 118, f.), it will be seen that its coast-lines and islands are almost slavishly adhered to on both maps. To this the Nancy map adds a Scandinavia, with Iceland, the east coast of Greenland, and a northern land connection between the latter and Russia. On the later map Scandinavia has been given a somewhat altered form, and Greenland has a west coast. The Nancy map has few names, many more being mentioned in the text, especially in Denmark. Even as regards Denmark they are evidently to a great extent taken from an older itinerary like that of Bruges ["Itineraire Brugeois," cf. Storm, 1891, p. 19]. Some of the names on the map, like "bergis," "nidrosia," etc., may be taken from older compa.s.s-charts; both texts have the northern form "Bergen." Headlands, bays and islands (on the coasts of Norway, Iceland and Greenland), for which he had no names (and which moreover are due to the free imagination of the draughtsman), have been designated in the Nancy text by Latin numerals ("Primum," "Secundum," etc.), or are simply named after each other (in Iceland), a sure sign that Clavus neither knew nor had heard anything about these coasts.

[Ill.u.s.tration: Copy, of about 1467, of Claudius Clavus's later map. The copy was executed by Nicolaus Germa.n.u.s. Owing to the map being transferred to the latter's trapezoidal projection, with converging meridians, Greenland, for instance, has been given a very oblique appearance]

[Sidenote: Mystification in Clavus's geographical names]

On his later map Clavus has made up for the want of names in an astonishing way. On some of the coasts he has continued to use Latin numerals for bays, etc., but side by side with this on the sh.o.r.es of the Baltic and in Sweden he has used Danish numerals, such as, "Forste aa fluuii ostia" (First river, river-mouth), "Anden aa" (Second river) ..., etc. The southerners, who did not understand Danish, of course regarded these as names, and subjected them to all sorts of corruptions. Matters became worse when in Gotland and Norway he used as the names of headlands and rivers the words of a meaningless rigmarole: "Enarene," "apocane,"

"uithu," "wultu," "segh," "sarlecrogh," etc. (evidently corresponding to children's rigmaroles like "Anniken, fanniken, fiken, foken," etc.)[245]

In Iceland he used the names of the runic characters for headlands and rivers; but most remarkable of all are his names in Greenland, alternately for headlands and the mouths of rivers(!). If, as shown by Bjornbo and Petersen, these are read continuously from the most northern headland on the east coast round the south of the country, the following verse in the dialect of Funen is the result:

"Thaer boer eeynh manh secundum [== ij ?][246] eyn Gronelandsz aa, ooc Spieldebedh mundhe hanyd heyde; meer hawer han aff nidefildh, een hanh hawer flesk hinth feyde.

Nordh um driuer sandhin naa new new."

(There lives a man (in ?) a Greenland river, and Spieldebedh is his name; he has more vermin (?) than he has fat bacon, etc.)

The verse, as pointed out by Axel Olrik, is evidently an imitation or travesty of the folk-songs, and, as Karl Aubert has shown,[247] its prototype must certainly have been the first verse of the same folk-song that is now known in Sweden by the name of "Kung Speleman":

"Dher bodde een kjempe vid Helsingborg, Kung Speleman mnde han heta, Visst hade han mera boda solf, an andra flesket dhet feta.

Uren drifver noran, och hafvet sunnan for noran."

(There lived a giant by Helsingborg, King Fiddler was his name.

Sure he had greater store of silver Than others of fat bacon, etc.)

This method of fabricating geographical names adopted by Clavus recalls the designation of the notes in the mediaeval scale, for which the words of a Latin hymn were used, and it seems likely that this is what he has imitated. But his mystification, with all these strange names which no one in Southern Europe understood, and which in course of time underwent many corruptions, has caused a good deal of trouble; many intelligent men have racked their brains to discover learned etymological interpretations of their origin, until Bjornbo's lucky find of the later text of Clavus solved the riddle.

[Sidenote: Different views of Clavus's maps and their origin]

Bjornbo and Petersen, who by their valuable work on Claudius Clavus with a reproduction of this text have the credit of throwing light on the relation between his first and second maps, have put forward the view that Clavus must have made his first map (the Nancy map) with its Latin text in Italy; but curiously enough they think he entirely rejected the Italian compa.s.s-charts as unsuitable for the representation of the North, and constructed his delineation of the northern regions independently of them, as an addition to Ptolemy's coast-lines, simply from information he had derived from northern sources. After this we are to suppose that, in order to extend his geographical knowledge, he went back to Denmark; and since the authors place reliance on Clavus's a.s.sertion (in his later text) that he had seen the places himself, they even credit him with having made a voyage of geographical exploration, first to Norway (Trondhjem) and then to Greenland. And then he is supposed to have drawn his later map, and written the text for it (in Latin), in the North.

I have come to an entirely different conclusion. His older map must be based, in my opinion, not only on Ptolemy, but to a great extent on Italian maps. His later map and text, I consider, show beyond doubt that he cannot have been either in Norway or Greenland, and I cannot find a single statement in the Vienna text, or any coast-line in his later map, which shows that he was outside Italy in the period between the two works.

Doubtless the delineation of Denmark, especially Sealand, is more detailed in the second map; but the additions do not disclose any more local knowledge than might be attributed to Clavus as a native of Funen before his first map was drawn, even though he had not then ventured to change the form of Ptolemy's Scandia, which to him, of course, became Sealand.

After this first attempt, however, he may have gained courage to launch out further with his knowledge. He may also have discovered a few fresh pieces of information, in the papal archives, for instance. Besides this, he may, of course, have received oral communications from people from the northern countries; but even of this I am unable to find sure signs. In consideration of the imaginative tendencies shown by Clavus in his distribution of names, and to some extent in the coast-lines on his map, which perhaps may also have a.s.serted themselves in his statement that he had seen a complete MS. of Livy in Soro monastery,[248] we shall scarcely be insulting him if we believe his statements (in two pa.s.sages of the Vienna text) that he himself had seen Pygmies from a land in the North, and Karelians in Greenland, to be rhetorical phrases, calculated to strengthen the reader's confidence, and to mean at the outside that he had seen something about these people in older authorities.

After having heard my reasons, Bjornbo and Petersen have in all essentials come round to my views. In particular they agree with me that Clavus cannot have been in Greenland, but that the delineation of that country on his later map is based on the Medicean map of the world, which will be mentioned later. I therefore consider it superfluous to combat any further here the reasons given in their work for their former view.

Claudius Clavus's task must have been to supplement the newly discovered atlas of Ptolemy by what he knew of the North; and to this end his maps were drawn, either by himself or by a professional draughtsman in Italy from his instructions. The text was prepared after each of the maps, as a description of it; and the lat.i.tudes and longitudes are taken from the map [cf. Bjornbo and Petersen, 1904, p. 130]. With the superst.i.tious respect of the period for older learned authorities in general, and for Ptolemy in particular, he did not venture to alter the latter's coast-lines or lat.i.tudes as far as they extended; even in the Danish islands he has done so with hesitation, thus Sealand in his first sketch [the Nancy map] has still the same form as Scandia in Ptolemy, etc. He then added to the latter's coast-lines what he knew or could get together from other quarters.

[Sidenote: Sources and genesis of the Nancy map]

His first map [the Nancy map] may presuppose the following sources, besides Ptolemy's various maps of Northern Europe; Pietro Vesconte's mappamundi (circa 1320) in Marino Sanudo's work,[249] and the anonymous mappamundi, now preserved in the so-called Medicean Marine Atlas, of 1351, at Florence.[250] In addition to these, either the Bruges itinerary itself [Itineraire Brugeois, cf. Storm, 1891, p. 19], or one of its earlier sources. Possibly he also had, in part at all events, a tract [in Icelandic ?] that is included in the fourth part of the "Rymbegla" [1780]; that he also knew of the Icelandic sailing directions, as a.s.sumed by Bjornbo and Petersen, I regard as less certain, although not impossible; perhaps it would be safer to suppose that he may have seen some statements from Ivan Bardsson's description of Greenland, in an itinerary, for instance. I have not been able to find any certain indication of his having been acquainted with the Icelandic geography mentioned on p. 237; perhaps he may rather have known of the land connection between Greenland and Russia from some tale or other, or from a legendary saga;[251] from the same source (or from Ivar Bardsson's description ?) may also be derived the name Nordbotn (cf. p. 171, note 1), which is not known in the Icelandic geography, but which seems most probably to be a legendary form.

Certain names, such as those of the bishops' sees in Norway and Iceland, Clavus may easily have found in the papal archives in Rome.

In the first place, exactly following Ptolemy, the draughtsman has marked Ireland with the islands around it and six Hebrides to the north-east, Scotland with the island of Dumna and the archipelago "Orcadia" to the north (the island of Ocitis a little farther east), and the south coast of Thule farther north; next Jutland with its small islands round about, and with the large island of Scandia, which, of course, became Sealand (he has added Funen and a number of other islands); finally the coast of Germany and Sarmatia eastwards to 63 N. lat., and with the same number of river-mouths as in Ptolemy. As this coast does not extend nearly so far to the east as does the Baltic on the compa.s.s-charts, it resulted that Clavus's Baltic became much shorter than that of the charts, and its shape had to be altered to suit Ptolemy's coast-line. Then, at its northern end, the draughtsman has placed possibly Pietro Vesconte's Scandinavian peninsula, going out towards the west (see the two maps, pp. 223, 224); but as he saw Norway on the compa.s.s-charts extending west as far as to the north of Scotland, where on Ptolemy's map he found Thule, it was natural that he should take the latter to be the southern point of Norway, and he was obliged to move Vesconte's peninsula farther to the west. Its south coast may have been drawn with the Medici map, or a similar one, as model.

As the southern coast of the Baltic was moved far to the south, after Ptolemy, and Jutland was given a different and smaller form than on the Medici map, besides a marked inclination to the east, and as Skne had to be near Sealand (Scandia), the draughtsman was obliged to move the peninsula corresponding to Skne about five degrees to the south. The south coast of the peninsula on the north of Scotland on the Medici map (see pp. 236, 260) corresponded very nearly to the south coast of Thule (with an east-south-easterly direction) on Ptolemy's map; it lay in an almost corresponding lat.i.tude, but on account of the puzzling prolongation of Scotland to the east on Ptolemy's map, it had to be moved a good fifteen degrees of longitude to the east. Thule was thus united to Norway[252] and its south coast was given exactly the same shape as the south coast of the peninsula in question, with three arched bays (the broadest on the east) and a projecting point towards the south-east. The coast between this promontory and Skne may then have been drawn with the same number of four large bays as on the Medici map: a deeper one farthest west, then a broad peninsula, next two wide, open bays, with a narrow peninsula between them, and finally a smaller bay opposite Sealand. The "Halandi" of the Nancy map is thus brought to the corresponding place with the "Alolanda" of the Medici map (p. 236).[253]

Thus far it may be fairly easy to compare the maps; but then Norway according to most of the compa.s.s-charts ought not to have any considerable farther extension to the west, while on the other hand Northern ideas demanded a Greenland in the far west, as well as a land in the north between that and Russia. With the latter the westernmost tongue of land in Norway on the Medicean mappamundi[254] agrees remarkably well. The southern point of Clavus's Greenland has also the same length in proportion to the west coast of Ireland, and about the same breadth, as on this map. There was also an extensive ma.s.s of land in the north. According to various representations, such as those of Vesconte's mappamundi, Saxo's description (cf. p. 223), and others, there should be a gulf on the north side of the Scandinavian Peninsula. According to representations like that of the Lambert map at Ghent (cf. p. 188), this arm of the sea had the same form as that on the south side of Scandinavia, and there should only be a narrow isthmus between these two arms of the sea, connecting the peninsula with the mainland (cf. Saxo). On the Nancy map, too, the north coast of Scandinavia is drawn almost exactly like the south coast, with the same number of promontories and bays, which correspond very nearly even in their shape. In this way Clavus's "Nordhindh Bondh" [Norrbotn], also called "Tenebrosum mare" [i.e., the dark sea] or "Quietum mare" [the motionless sea], may have originated. This remarkable bay is connected on his map with the Baltic by a ca.n.a.l (which is also mentioned in the Vienna text). By this means Scandinavia really becomes an island. Clavus cannot have acquired such an idea from any known source, although, as already mentioned, Saxo says that it is nearly an island (p. 223); but similar conceptions seem to have arisen in Italy (cf. above on Pietro Vesconte's mappamundi, p. 223).

[Ill.u.s.tration: Scandinavia on the map of Europe in the Medici Atlas (of 1351). The scales of lat.i.tude and longitude are here added from Ptolemy's maps. The network of compa.s.s-lines is omitted]

The south coast of Norway [with "Stauanger"] and the southern point of Greenland retained on Clavus's map the same relation of lat.i.tude, a difference of 1-1/2, as the corresponding localities on the Medici map, with very nearly the same degrees of lat.i.tude as on the latter, if we there employ a scale of lat.i.tude calculated upon this map's representation of Spain (the Straits of Gibraltar) and France (Brittany), and use Ptolemy's lat.i.tudes for these countries. This has been done in the reproduction of the Medicean mappamundi on p. 236.[255] The scale of longitude is calculated in the same proportion to the lat.i.tude as in Ptolemy. In some tract like that included in the fourth part of the "Rymbegla" [1780, p. 466] Clavus may have found that Bergen lay in lat.i.tude 60 and so placed the town on the west coast of Norway in this lat.i.tude according to his own scale (on the right-hand side of the Nancy map, see p. 474). In relation to the south coast of Norway Bergen was thus brought 3/4 farther south than "c. bergis" on the Medici map (above).

Calculated according to Ptolemy's scale of lat.i.tude (on the left-hand side of the Nancy map), Bergen was consequently placed in Clavus's text in 64, while the southern point of Greenland is placed in 63 15',[256] a difference in lat.i.tude of 45' (in the Vienna text the difference is 35'), while in reality it is 38'; a remarkable accidental agreement. According to Clavus's own scale of lat.i.tude on the right-hand side of the Nancy map, we get the following lat.i.tudes: Bergen 60, the southern point of Greenland 59 15', Stavanger 58 30'. In reality the lat.i.tudes of these places are: 60 24', 59 46', and 58 58'. This agreement is remarkable, as a displacement of the scale of lat.i.tude half a degree to the north on the Nancy map would give very nearly correct lat.i.tudes.[257] The mutual relation between the lat.i.tudes of the three places may, as we have seen, be explained from the Medici map, but hardly from a possible acquaintance with the Icelandic sailing directions; for according to these Bergen and the southern point of Greenland would be placed in the same lat.i.tude, since we are told that from Bergen the course was "due west to Hvarf in Greenland."[258] The Medici map may also give a natural explanation of places like Bergen and the southern point of Greenland having been given by Clavus a lat.i.tude so much too northerly (even in the Nancy map), and of the southern point of Greenland having only half a degree more westerly longitude than the west coast of Ireland.[259]

Iceland lay, according to the Bruges itinerary, midway between Norway and Greenland, precisely as on the Nancy map. Between Norway and Iceland, according to the same itinerary, lay "Fareo" [Faero], and the fabulous island "Femoe," "where only women are born and never men."

After speaking of the "third headland" in 71 on the east coast of Greenland, the Nancy text goes on:

"But from this headland an immense country extends eastward as far as Russia. And in its [i.e., the country's] northern parts dwell the infidel Karelians ('Careli infideles'), whose territory ('regio') extends to the north pole ('sub polo septentrionalis') towards the Seres[260] of the east, wherefore the pole ['polus' == the arctic circle ?], which to us is in the north, is to them in the south in 66."

It is probable, as suggested by Bjornbo and Petersen, that these "Careli infideles" are identical with those who are found almost in the same place, in the ocean to the north of Norway, on one of the maps in Marino Sanudo's work (in the Paris MS., see above, p. 225), and who on other maps belonging to that work are placed on the mainland to the north-east of Scandinavia. As pointed out by Storm, "Kareli" are also mentioned together with Greenland and "Mare Gronlandic.u.m" in the Bruges itinerary.

Bjornbo and Petersen maintain that Claudius Clavus has here consciously put forward a new and revolutionary view which was a complete break with the cosmogony of the whole of the Middle Ages, since according to the latter the disc of the earth was entirely surrounded by sea to the south of the North Pole, as represented on the wheel-maps. I think this is attributing to Clavus rather too much original thought, of which his maps and text do not otherwise give evidence. It is, of course, correct that the idea of land, and inhabited land, too, at the North Pole, or to the north of the Arctic Circle, did not agree with the general learned conception of the Middle Ages; but the same idea had already been clearly enough expressed in Norwegian-Icelandic literature. Even the Historia Norwegiae has inhabited land beyond the sea in the north, and the Icelandic legendary sagas and Saxo have it too. In addition to these, the tract included in the "Rymbegla" says distinctly (see above, p. 239) that this land in the opinion of some lies under the pole-star (cf. Clavus's expression: "sub polo septentrionalis"). The fact that the continent on the Medicean map of the world extended boundlessly on the north into the unknown (whereas Africa ended in a peninsula on the south) must have confirmed Clavus in the view that the land reached to the pole. To this was added, what perhaps weighed most with him, the fact that such a view did not conflict with Ptolemy, whose continent also had no limit on the north.

On the connecting land in the north is written, on the Nancy map: "Unipedes maritimi," "Pigmei maritimi," "Griffonii regio vastissima," and "Wildhlappelandi." As these names are not mentioned in Clavus's text, it is uncertain whether the fabulous creatures may not be to some extent additions for which he is not responsible.

After the map was drawn, with its bays and headlands, and the coast of Scandinavia provided with a suitable number of islands, Claudius Clavus set himself to describe it; where he had no names from earlier sources, he numbered the headlands, bays and islands, "Primum," "Secundum," etc.

A remarkable thing about the Nancy map is that it has two divisions of lat.i.tude: one according to Ptolemy on the left-hand side of the map, and another according to Clavus himself, on a scale four degrees lower, on the right-hand side. According to the latter, Roskilde would have a longest day of seventeen hours (through a transposition the Nancy map gives seventeen hours thirty minutes), which, as pointed out by Bjornbo [1910, p. 96], exactly agrees with what Clavus may have learnt from a Roskilde calendar ("Liber daticus Roskildensis") of 1274. Bjornbo has also remarked that Bergen is given a remarkably correct lat.i.tude, 60 (the correct one is 60 24'), and thinks it possible that there may have been a Bergen calendar which Clavus has used. But a more likely source, unnoticed by Bjornbo, is to be found, as mentioned on p. 260, in the "Rymbegla" tract, where the lat.i.tude of Bergen is given as 60. It is true that the same tract gives the lat.i.tude of Trondhjem (Nidaros) as 64, which does not agree with the Nancy map, where there is a difference of only 2 between Bergis and Nidrosia. Even though it is probable that Clavus was acquainted with some such tract, with which his statement as to land at the North Pole also agrees, it may have been a somewhat different version from that which found its way into the "Rymbegla," and perhaps the lat.i.tude of Trondhjem was not mentioned there. On the other hand, he may have found, there or elsewhere, the lat.i.tude of Stavanger given, 1-1/2 farther south than Bergen (?).

If we a.s.sume that Clavus, even in the construction of his first map, made use of the Medicean map of the world, and that his Greenland is the most westerly peninsula of the latter's Norway, it will seem strange that he did not also draw the west coast of that peninsula, which would naturally become the west coast of Greenland. It is true that the Nancy map is only a copy, but as the west coast of Greenland is not mentioned in the copy of Clavus's text either, we are bound to believe that he did not include it. The margin on the western side of Clavus's first map was evidently determined by that of Ptolemy's map of the British Isles, and follows precisely the same meridian. Thus there was no room for the Medici map's peninsula corresponding to Clavus's Greenland. As already stated, it is difficult to get away from the belief that the Medici map was used for the east coast of Greenland, the south coast of Norway, etc.; the resemblances are too great, and otherwise inexplicable (cf. p. 261, note 3).

[Sidenote: Clavus's later map and text, and their genesis]

After the first map was drawn, Clavus may have made further cartographical studies in Italy, and may thus have become acquainted with other compa.s.s-charts, especially those of the Dalorto type. At the same time he may have obtained a new and more accurate determination of the lat.i.tude of Trondhjem, probably by the length of its longest day. As Trondhjem was an archbishopric, it is not unlikely that he found such a piece of information in the papal archives at Rome. He may then naturally have wished to bring his map more into agreement with his new knowledge, and this may have led to his later map, which is now known to us through several somewhat varying copies. To this he then wrote a new text (the Vienna text), which in all important points resembles the former, but has various additions and alterations. The later map has not the double scale of lat.i.tude on any of the copies known, but curiously enough only Ptolemy's degrees. Besides a more accurate delineation of Jutland and the Danish islands, especially Sealand, Bornholm and Gotland are drawn in closer resemblance to the Medici map; the south coast of Scandinavia has been altered to agree more with compa.s.s-charts of the Catalan type. In particular the south coast of Norway has been given the four characteristic promontories (as on the Dalorto map of 1339, and on the Modena map, etc.; cf. the reproductions, pp. 226, 231), and Bergen ("Bergis") has been placed at the head of the westernmost of the three bays thus formed, which is also a peculiarity of the maps of this type (the Catalan chart of 1375 has five promontories with four bays, cf.

Nordenskiold, 1896, Pl. XI.). The other two diocesan towns, Stavanger and Hamar, are placed at the heads of the other two bays to the east, and Stavanger has thus lost the remarkably correct position in relation to Bergen and the south point of Greenland which it had on the older map.

Trondhjem has been placed at the extremity of the westernmost promontory, possibly because there had been found a more correct determination of the lat.i.tude of the town, which was to be fitted into Ptolemy's graduation; thereby the shape of Norway has become still narrower and farther removed from reality.

From the "lac scarsa" (Lake Skara, i.e., Vener) with its river is derived the great lake "Vona" (Vener) in the centre of Scandinavia on all the copies of Clavus's later map, from which the river "Vona" (also mentioned in the Vienna text) runs into the deep bay by "Aslo" (Oslo) and the island of "Tunsberg." A connection, especially with Dalorto's map of 1339, seems again to be implied by Clavus's statement in the Vienna text that on Lister Ness "white falcons are caught" ("Liste promontorium, ubi capiuntur falcones albi"). On Dalorto's map there is a picture of a white falcon on the headland to the west of that which Clavus has made into Lister, and the words "hic sunt girfalcos" (here are hunting falcons). That Clavus has moved the hawks to a headland farther east is of small importance. Either he may have taken his hawks from Dalorto's or a similar map, or else they are derived from an older common source.

Through the alteration of the south coast of Norway, it became necessary to separate it from Thule, which again became an island as originally in Ptolemy; but on the copies of the map it has in addition the name "Bellandiar," which may be a corruption of Hetlandia (Shetland). The north-west coast of Norway has also been given a form which agrees better with the compa.s.s-charts, although it has a much more east-north-easterly direction than even on the Modena map; but this was, of course, necessary to make room for the sea "Nordhenbodnen" (Nordbotn). That the compa.s.s-charts might lead to something resembling Clavus's last form of Scandinavia, and especially of the south coast of Norway, is shown by the map of Europe in Andrea Bianco's atlas of 1436, which must have been drawn without knowledge of Clavus's work. If on this map we move the coast of the Baltic farther south and Skne also, which would be necessitated by a better knowledge of Denmark (and by the alteration of the map following Ptolemy), and draw the coast-line of Norway towards the east-north-east from the south-western promontory (instead of making it go in a northerly direction), we shall get a Scandinavia of very similar type to that in Clavus's later map.

[Ill.u.s.tration: The north-western portion of the map of Europe in Andrea Bianco's atlas of 1436. The compa.s.s-lines are omitted]

Bjornbo and Petersen have maintained in their monograph that Clavus must have been in Norway before he drew this map, and that amongst other things his remarkably correct lat.i.tude for Trondhjem must be due to his own observation of the length of the day at the summer solstice. Storm [1889, p. 140] seems also to have supposed that Clavus may really have been in Norway. To me it appears that his map and text are conclusive evidence against his ever having been there; for a man who had sailed to Trondhjem along the coast of Norway could not possibly have produced a cartographical representation of the country so entirely at variance with reality as Clavus has done, however ignorant we may suppose him. The fact in itself that "Trunthheim" (Trondhjem) or "Nedrosia" is placed at the extremity of the south side of the south-western promontory of the country is extraordinary. If he had come there asleep he could not have got any such idea; and for a man who had sailed in through the long channel of the Trondhjem fjord up to the town it is incredible. It is equally incredible that a man who had sailed along the coast from Stavanger and Bergen to Trondhjem could place the latter town in a lat.i.tude 10' to the south of Bergen, and only 10' to the north of Stavanger. We are not justified in attributing to Clavus such an entire lack of power of observation, especially if we are to suppose him capable of determining with remarkable accuracy the length of the longest day at Trondhjem. That Trondhjem is placed to the west of Bergen and Stavanger, that the Dovrefjeld is called a high promontory, while on the Nancy map it was inland, that Hamar ("Amerensis") is put on the sea-coast, etc., all shows the same want of knowledge of the country and its configuration. The names he may have taken from an itinerary or other sources, and, as already suggested, it is not unlikely that he may have found in the papal archives a fairly correct statement of the lat.i.tude (or length of the longest day) of Trondhjem, which was an archbishop's see. That the towns he gives are just those that are the heads of dioceses is perhaps an indication of a connection with the Vatican.

Clavus tells us further that

"Norway has eighteen islands, which in winter are always connected with the mainland, and are seldom separated from it, unless the summer is very warm," and that "'Tyle' [Thule] is a part of Norway and is not reckoned as an island, although it is separated from the land by a channel or strait, for the ice connects it with the land for eight or nine months, and therefore it is reckoned as mainland. The same applies to the sea 'Nordhinbodnen' [Nordbotn], which separates 'Wildlappenland' from 'Vermenlandh'[261] and 'Findland' by a long strait, since the countries are united by almost eternal ice."

This discloses an extraordinary lack of knowledge of Northern conditions.

Such a connection of the islands with the mainland by ice occurs, of course, nowhere on the whole outer coast of Norway from Faerder to the Murman Coast. On the other hand, the Gulf of Bothnia and the land archipelago are frozen over for a long time in winter, and it might be supposed that Clavus had heard reports of this. But I have not been able to discover any source from which he may have derived these fables. Most probably they are embellishments of the same kind as the eighteen islands of Norway, that form an arbitrary decoration of the coast-line of his map, a circ.u.mstance which does not hinder him from describing them as real.

Clavus has used the ice as a transition between the representation of his older map, where Thule was part of the mainland, and that of the later one, where it was made into an island.