In a previous work [*] I have attempted to show that the discovery of Arnhemsland must beyond any doubt be credited to the voyage of the yacht Arnhem, commanded by Van Colster or Van Coolsteerdt, which took place in 1623.  Since the Journal and the charts of this voyage are no longer available, we are without the most important data for determining with certainty between what degrees of longitude the Arnhemsland then discovered was situated.  To westward of it must be sought Van Diemens-and Maria’s-land, touched at in 1636 by Pieter Pieterszoon with the ships Cleen Amsterdam and Wesell) [**].  There can be no doubt that Pieterszoon must have sailed far enough to westward to have passed Dundas Strait, and to have reached the western extremity of Melville Island (Roode hoek = red point).  He took Dundas Strait to be not a strait, but a bay, and accordingly looked upon Melville Island not as an island, but as a portion of the mainland (Van Diemensland) [***].

[* See my Life of Tasman, pp. 100-102, and the Documents under No.  XIV, 2 infra.]

[** See the Documents under No.  XXV.]

[*** Maria-land lies immediately to eastward of Van Diemens-land, and to westward of Arnhems-land.]

In the course of these two voyages of 1623 and 1636, therefore, the whole of the north-west coast from Melville Bay to Melville Island was surveyed by Dutch ships.  But in the absence of charts made on these voyages it is impossible for us to say with certainty, whether the coastline can have been traced with correctness.  On this point also more light is thrown by the well-known chart of 1644, in which the results of Tasman’s voyages are recorded.  Tasman sailed along the whole of the coast, but in this case too, his observations were not on all points accurate.  Thus the situation of Wessel-eiland and the islets south of it, with respect to the mainland, is not given correctly by him; nor has he apprehended the real character of Dundas Strait and of Van Diemen’s Gulf, so that also according to him Melville island forms part of the mainland.  But for the rest Tasman’s chart also in this case approximately reproduces the coast-line with so much correctness, that we find it quite easy [*] to point out on the maps of our time the results of the Dutch voyages of discovery in this part of the Australian coast.

[* Chart No. 14 below may also be of excellent service here.]

Far more accurate, however, than Tasman’s chart is the chart which in 1705 was made of the voyage of the ships Vossenbosch, de Waijer and Nova-Hollandia, commanded by Maarten van Delft [*].  This chart may at the same time be of service to elucidate Tasman’s discoveries and those of his predecessors.  It is to be regretted, therefore, that it only embraces a comparatively small portion of the north-west coast, namely the part extending from the west-coast of Bathurst island and the western extremity of Melville island to the eastern part of Coburg peninsula and Croker-island.  This time again the real character of Dundas Strait and Van Diemens Gulf were not ascertained [**].