The Java analysis
Initial comparison had suggested that a series of new names began to appear along Java's north coast at about the time of the first Dutch voyage (1595-7). But were they really new or were they perhaps revivals; or were we dealing instead with different series of names in use at the same time?

To confirm the significance of this tentative finding beyond any doubt, an intensive examination was made of charts, maps and globes over the long period 1550-1650. By extending the study beyond the limits of strict relevance a wider picture was revealed and confident conclusions were possible within the chosen period. During the course of this, at least 1,500 Java names were noted. By a ruthless process of compression this vast number was reduced to those sixty-two names which had been included for the first time on maps produced between 1550 and 1620. The period after 1620 is dominated by the entirely new range of names displayed on the charts of Hessel Gerritsz. Since Llewellyn has none of these it seemed sensible to curtail the secondary investigation at this point. Under one head were included all the variant forms of any one name (however much they differed) when they seemed intended for the same place. The incidence of these names on the maps under study could have been arranged graphically at this stage but the scale and complexity would have buried any patterns there might have been. So this analysis was, in turn, put back into the cauldron to distil off still further its essential features. As we were, at this point anyway, concerned only with the presence or absence of a name, not with its spelling, nor with the order in which a sequence of names along the coast was presented on different charts, the first task was to provide each with the date of its first and last appearance. Those that continued throughout the period could be safely ignored and attention focussed instead on those that had been inserted or abandoned.

The result of this survey was a definite confirmation that an entirely new series of Java names began to appear about 1598. These names occur first on a detailed chart of Java, Sumatra and southern Borneo, entitled Nieuwe caerte op Java geteeckent. It was compiled by 'G.M.A.' [i.e. Willem] Lodewijcksz, engraved by Baptista à Doetechum and published at Amsterdam by Cornelis Claesz, probably in 1598. { There is some doubt whether this detailed map was an integral part of the book but f.24v of the French edition includes the note 'Icy doibt estre mis la Carte de Iava & Sumatra', which must surely refer to this. It must be noted, though, that, with the exception of Jacatra, none of the toponymic innovations of the Lodewijcksz map have been detected in the book's text}. Examination of some twenty-five earlier maps compiled in the period up to and including 1598 revealed a consistent total of between fifteen and nineteen Java names. That the greatest number detected, twenty-three, had appeared on a map of c 1540 illustrates the static nature of East Indies cartography in the 16th century. The selection of names was fairly constant too. The forty-nine names that are found on Lodewijcksz's chart would therefore have provided a striking contrast to this traditional, unchanging picture had they merely combined different earlier selections. In fact, two-thirds of Lodewijcksz's total are innovations and his chart thus marks a transfer of initiative from the Portuguese to the Dutch which is, in a cartographic sense, as dramatic and sudden as the changes that were taking place in the geopolitical sphere.

Lodewijcksz sailed with Cornelis de Houtman as supercargo on the first voyage made by Dutch ships to the East. This fact alone provides a strong hint that the new names were gathered then and made known after the expedition's return in the summer of 1597. This is confirmed by Isaak Commelin's account of this voyage, published in Dutch in 1646 and made available in English in 1703. (It is to this latter version that references are made). Commelin's account is quite different from the one issued by Barent Langenes, a few months after the expedition's return. It is also twice as long. Where Langenes had mentioned only one of Lodewijcksz's innovations (and that the most obvious, Jacatra), Commelin refers specifically to fifteen of the thirty-two Java names that Lodewijcksz introduces.

Their source was probably a Portuguese living in Java, Pedro de Tayda [Truide, Taydo, Tayde or de Ataide], whom they met in Bantam (and who is first mentioned on 25 July 1596): "a famous Pilot, who had frequented all the coasts and Islands of the East Indies and made Maps of them all, which he promis'd to shew the Dutch. This gave them great Hopes of discovering more of that Country, than he had discover'd to them before" (Commelin p.156). Hakluyt gives the following account by Barent Langenes: "Among the Portingalles there was one that was borne in Malacca, of the Portingalles race, his name was Pedro Truide, a man well seene in trauayling, and one that had beene in all places of the world". De Tayda was murdered three weeks later (16 August) apparently as a direct result of the sharing of his cartographic knowledge with the Dutch. Presumably his maps and pilotage instructions, or copies of them, were taken back to Holland where they would have been gladly received by the publisher, Claesz - much as Bartolomeu Lasso's atlas had been a few years previously. Petrus Plancius, referred to de Tayda in 1598 and in 1599 stated that "it is therefore highly necessary that the masters and commandants read and well consider the writing of Pedro de Tayde and other sailors" [PMC 4:3].

{An alternative explanation would be that the maps brought back by Houtman were compiled in Java, in the course of the numerous discussions de Tayda had with the Dutch. The title of Lodewijcksz's map ends '...delineata in insula Iava, ubi ad vivum designantur vada et brevia scopulique interjacentes descripta a G.M.A.L.' [...was drawn on the island of Java, where the shoals and shallows and intervening cliffs are marked out from life, described by G.M.A.L] - with thanks to Mary Pedley for help with this translation.}

An independent Portuguese source helps to explain simultaneous additions to the toponymy of southern Borneo, not visited by Houtman's ships [Broek]. It also shows why the new names are not Dutch words, which would stand out clearly from the earlier ones, but rather a new selection of indigenous or Portuguese names. However, the inclusion on most maps of the early 17th century of some at least of Lodewijcksz's innovations justifies us in treating them as one of the most important, if not the most important, element in the cartographic context we are attempting to establish. It has not, apparently, been noted before. When considering the dating of charts supposed to have been produced at the end of the 16th century, the inclusion of Lodewijcksz's Java names, therefore, must point to 1597, or more realistically 1598, as their earliest possible date. Llewellyn's atlas includes a number of these names. Indeed, it has one of the largest concentrations of these so far identified. [See the Table of significant Java names].

In a comparable study on the maps of Borneo, already mentioned, Broek [p.148] cites only five maps for the period between Lodewijcksz and Gerritsz's much improved charts of the 1620s, and three of these are versions of Plancius's large world map of 1592, and thus in their content earlier than Lodewijcksz. In the present investigation thirty-four charts, maps and globes of this same period, of sufficient scale to offer at least ten Java names, have so far been identified and examined.

When judged purely quantitatively, the density of Java names in Llewellyn's atlas is only matched by two others: Gabriel Tatton's chart of the region between Bengal and Florida (Biblioteca Nazionale, Florence), provisionally dated to 1600, and Willem Blaeu's wall-map of Asia, one of a series of the continents first produced in 1608. All three contain roughly two-thirds of the thirty-two names introduced by Lodewijcksz. Blaeu's map has in addition a further five names not found earlier, which represent the only significant Java innovations between 1598 and 1620. The new 'Blaeu' names were not known to Llewellyn. After that date Lodewijcksz's names are largely ousted by Gerritsz's and by 1657, when Janssonius issued a detailed chart of the island, only seven remained.

As an indication of possible refinements to these generalisations a few specific examples will have to suffice:

1. The most obvious name from the Dutch period is Jacatra. Formerly Sunda Calapa (as it was recorded by, e.g. Linschoten) and, after 1621, renamed Batavia, it remains, as modern Jakarta, a city of importance. Few maps after 1598 omit it, although none had previously named it.
2. Surabaja, another important town, first appears on Lodewijcksz's chart, as does Tegal.
3. Solpherburg, first included on Blaeu's map of 1608, is the earliest Dutch name given to a Javanese feature. The 1703 account of the first voyage describes "a great Mountain of burning Sulphur", at the east end of the island behind Panarukan, "which opened for the first time in the year 1586" (Commelin p.181).
4. The Straits of Sunda, that important channel dividing Sumatra from Java, must have been known to the Portuguese for a century and it was certainly named thus before the Dutch arrived. Yet no map before 1598, so far seen, names the strait and the first to do so is not Lodewijcksz's chart but an otherwise traditional map by Hulsius. Besides another map that Hulsius issued in 1602, the strait's name has not been seen again until the 1608 Blaeu map. From then onwards it becomes a standard feature. In this, as in so many other instances, it was apparent accident that determined the date at which a name was first absorbed into the cartographic bloodstream and imitation that kept it there.

The outline commonly given to Java was essentially indefinite, since its south coast was unknown. Van Linschoten, indeed, had even questioned whether Java was an island at all. Despite the fact that Houtman's fleet returned by way of the south coast - the first European vessels known to have circumnavigated the island - there was no measurable improvement in the island's shape. It was left to Gerritsz (evidently on his engraved chart of 1618) to replace this shapeless mass with an outline that is strikingly close to its real form.



So much for the context - how does Llewellyn fit into it?

When Llewellyn's version of Java is measured against all the features which the analysis had shown to be significant, it falls naturally into a slot close to the beginning of the Dutch period. Eight of Lodewijcksz's new Java names seem to disappear after 1600, yet all are to be found in Llewellyn's atlas. One special instance can be mentioned. Arosbaja, at the west end of Madura Island, is conveyed by Lodewijcksz as Rossumbaya. This is clearly a mistake and it was to be corrected by Blaeu in 1608. But Llewellyn and Tatton both repeat Lodewijcksz's error. These three charts are bound even closer together by the inclusion of three names not traced anywhere else: Chuconin, Labuan, Meleasseri [see the Table of significant Java names. See table on next page].



While Lodewijcksz's seems clearly to be the mother-chart that introduces the period of Anglo-Dutch dominance in the East, Llewellyn's picture of Java is far from being a slavish copy of it and several of the name forms are noticeably dissimilar in the two versions. Beyond that, Llewellyn includes one name not found on Lodewijcksz's chart, nor, apparently, on any other, Sigulo, just east of Jakarta. It is also significant that Llewellyn's Juama (for Lodewijcksz's Ivanna) is the spelling adopted in the 1703 Commelin account (p.197). Tatton's Java names can all be found, with minor variations, on Lodewijcksz's chart, and Blaeu's map of 1608 is even closer to the latter.

But the essential individuality of Llewellyn's work shows that we are dealing with an important new source. None of the contemporary Dutch charts that I have been able to examine, by Cornelis Doedtsz, Everts Gijsberts and the brothers Harmen and Marten Jansz, all members of the so-called Edam School of chartmakers, reflect to anything like the same degree the new generation of Java names, and some remain entirely ignorant of them.

There is not space to mention more than the barest details of parallel studies into other sections of Llewellyn's atlas. The greater part of the impact of Houtman's voyage (or perhaps more properly of de Tayda's knowledge) was expended on Java but ripples reached Borneo and Sumatra at least, adding a further four new names for the former and a possible eleven for the latter. Llewellyn includes all but one of these. In no case does his information seem to conflict with a 16th-century date.


Attention was also paid to the shape of Japan. In 1595 Ortelius had added to his atlas Luis Teixeira's picture of Japan, an immeasurable improvement on its grotesque predecessors. Manuscript charts, in particular, were slow to adopt this and it is therefore significant both that Llewellyn should favour Teixeira's form and that his distinctive variation of it precludes his having copied the engraved map. Llewellyn's Dutch contemporaries, meanwhile, persevered with the Vaz Dourado form of half a century earlier, as did Tatton.