PROCEEDINGS OF SECTION C. 405 



danger signals informing him of the proximity of coral reefs. He had 

 previously observed such islands in the Paumotus, but seems not to 

 have been sufficiently familiar with them to appreciate their significance 

 and profit by the warning they conveyed. 



The other type appears as a long array of hilly timbered islands 

 standing like sentinels ofi the Queensland coast. Here and there 

 in ones and twos, but often clustered in archipelagoes, they stretch 

 from Thursday Island, or, more accurately. Mount Cornwallis, in the 

 north to the Keppel Islands in the south. Rarely does a coaster lose 

 sight of one before another rises into view. Each island is associated 

 with the mainland nearest to it by identity not only of rock structure, 

 but also of fauna and flora. They impress a casual observer as being 

 a part severed from the main. We may dispense with arguments 

 that they are aught else and consider the manner of their severance. 



In other regions of the world — for example the Chilian coast — 

 a long stream of off-shore islands is held to indicate subsidence. This 

 is the first explanation, and we hold the true one, which these islands 

 suggested to geologists. Like a conflagration subsidence destroys its 

 own history, but a couple <"'( hundred feet would be a moderate estimate 

 of the descent which isolated these rock masses. Such a movement 

 would suffice to build much of the Barrier Reef on Darwinian lines. 



Professor Agassiz, who avoids this deduction, regards the high 

 islands as separated from the main, not by subsidence but by " erosion 

 and denudation." We cruised along the Barrier with his memoir in 

 hand and studied it in sight of the subject. Though impressed by his 

 vast experience and knowledge, we are unable to accept his interpre- 

 tation of this crucial point. 



Denudation fails before reaching base level; it could not reduce 

 a cape to an island. Marine erosion only takes effect upon a coast 

 fronting the open sea ; it could not excavate tortuous sheltered channels. 

 A well-known beauty spot of the Queensland coast is the Whitsunday 

 Passage. On either side extend an intricate dissection of mountainous 

 land seamed by waterways 100ft. deep. To this archipelago we especi- 

 ally point as a characteristic product of subsidence ; we submit that 

 its features could not be evolved by erosion and denudation. 



Professor Agassiz, if we understand him aright, traces the history 

 of the Barrier Reef without the intervention of subsidence. He con- 

 siders that this region has been stationary since the Cretaceous era, 

 that numerous continental islands similar to those which now fringe 



the coast have been cut down by marine erosion, and that on their 

 bases have spread as a " thin veneer " the reefs now existing. The 

 unfinished product of marine erosion might be imagined as a flat cake, 

 with a median or lateral undenuded residue, as shown in figure. Instead of 



