Analytical methods
Christ Church's Donors' Book (p.84) records the gift of the atlas (West Table A 3) by Llewellyn's sons, William and Martin, in 1634:

"Gulielmus lluellin & Martinus Filij Martini Civis londinensis D.D. Cartas Geographicas patris manu depictas & ab eodem observatas sinus Maritimi in Africa & Asia a Capite bonae Spei ad Regnum China & partem Americae peruanae" [William Llewellyn & Martin, sons of Martin Citizen of London, made the gift of the geographical charts, drawn in the father's hand and according to his own observations, of maritime straits in Africa & Asia, from the Cape of Good Hope to the Kingdom of China and towards South America]

Young Martin, who was later to become famous as a physician to Charles II and as a writer, was an undergraduate at Christ Church, possibly in 1634, certainly two years later, and went on to get his BA in 1640. But when was the atlas compiled? Christ Church's records cannot help there and to answer this we must turn to the atlas itself. But how should we begin; what is the accepted procedure for evaluating a whole atlas?

Discounting repeated names on overlapping sheets, Llewellyn's atlas contains some eleven hundred names, far too large a total to be systematically compared with similar columns of place-names from other supposedly comparable charts. And what was comparable anyway? Had Llewellyn died at a ripe old age we might have to study the entire period from, say, 1580 onwards, a span of half a century.

A search for obviously datable features met with no success. The maps of later centuries might reflect discoveries, new surveys or theories, changes in political boundaries, the foundation of cities, the impact of man on his environment; but the East in the late 16th and early 17th century was not responsive to tests of these kinds. True, the turn of the century saw the dramatic replacement of the Portuguese by the Dutch and English, but the presence of a new colonial power would not necessarily stand out on a chart.

What about visual comparison, then? R.A. Skelton warned of basing "a heavy load of theory on a visual impression" [Skelton, 1965 p.6]. Comparison of Llewellyn's outlines for the islands of the East Indies shows them to be at best vaguely similar, sometimes noticeably dissimilar, and never exactly the same as contemporary forms. Visual comparison tends, anyway, to judgements that are both subjective and imprecise.

Another possible point of departure would have been an examination of voyages to the East within the period we have delimited. Again a timely word of warning from Skelton: "Traces of voyages of which the written record is wanting may be found in the maps; and it is equally clear that the maps have suffered a high degree of wastage and loss. In other words, neither the series of voyages, represented by documents, nor the series of maps, represented by extant specimens, is complete" [Skelton, 1965 p.15]. Llewellyn's atlas contains no detected reference to any particular voyage and the spread of place-names is nowhere abnormally dense enough to suggest special knowledge. Since an analysis of Borneo toponymy by Broek confirmed the gap between the acquisition of information about the East Indies and its appearance on charts, this line of approach seemed unprofitable.

Instead it was decided to treat the problem as a strict exercise in comparative cartography. If analysis of all available maps and charts were to reveal certain clear patterns then it would not necessarily matter if the reasons for these remained obscure. They would provide the cartographic evidence we needed to establish a purely cartographic context, against which to evaluate Llewellyn's contribution. Never mind the mistakes and omissions which we, with the benefit of hindsight could detect; Llewellyn was a man of his time and presumably as informed or as ignorant as his contemporaries. Since this cartographic context had never previously been identified, this was clearly where the start had to be made.

A detailed analysis of all sixteen sheets in the atlas was obviously out of the question; there had to be some selection. The precise method hit upon owes something to the techniques of medicine. One area was chosen for intensive study and, rather like a doctor's blood sample, it was hoped that microscopic examination in the laboratory would show up traces of peculiarities infecting the whole. The essential difference, of course, is that the wider relevance of any cartographic findings would have to be checked; they could not, as in the medical analogy, be assumed. But hopefully the range of possibilities would be much reduced.

The area chosen was Java, partly because a pilot study revealed a development in its toponymy not paralleled elsewhere, and partly because its choice by the Dutch as the seat of their government in the East pointed to its special importance in the 17th century.