Pliny,[66] who published his work on Natural History about A.D. 77, states that the turtles of the Indian Sea are of such vast size that a single shell is sufficient to roof a habitable cottage, and that among the islands of the Red Sea the navigation is mostly carried on in boats formed from this shell. Ælian,[67] about the middle of the third century of our era, is more specific in his statement, and says that the Indian river-tortoise is very large, and in size not less than a boat of fair magnitude; also, in speaking of the Great Sea, in which is Taprobana (Ceylon), he says: "There are very large tortoises generated in this sea, the shell of which is large enough to make an entire roof; for a single one reaches the length of fifteen cubits, so that not a few people are able to live beneath it, and certainly secure themselves from the vehement rays of the sun; they make a broad shade, and so resist rain that they are preferable for this purpose to tiles, nor does the rain beating against them sound otherwise than if it were falling on tiles. Nor, indeed, do those who inhabit them have any necessity for repairing them, as in the case of broken tiles, for the whole roof is made out of a solid shell so that it has the appearance of a cavernous or undermined rock, and of a natural roof." El Edrisi, in his great geographical work,[68] completed A.D. 1154, speaks of them as existing down to his day, but as his book is admitted to be a compilation from all preceding geographical works, he may have been simply quoting, without special acknowledgment, the statements given above. He says, speaking of the Sea of Herkend (the Indian Ocean west of Ceylon), "It contains turtles twenty cubits long, containing within them as many as one thousand eggs." Large tortoises formerly inhabited the Mascarene islands, but have been destroyed on all of them, with the exception of the small uninhabited Aldabra islands, north of the Seychelle group; and those formerly abundant on the Galapagos islands are now represented by only a few survivors, and the species rapidly approaches extinction. I shall close this chapter with a reference to a creature which, if it may not be entitled to be called "the dragon," may at least be considered as first cousin to it. This is a lacertilian of large size, at least twenty feet in length, panoplied with the most horrifying armour, which roamed over the Australian continent during Pleistocene times, and probably until the introduction of the aborigines. Its remains have been described by Professor Owen in several communications to the Royal Society,[69] under the name of _Megalania prisca_. They were procured by Mr. G. F. Bennett from the drift-beds of King's Creek, a tributary of the Condamine River in Australia. It was associated with correspondingly large marsupial mammals, now also extinct.