180 



SCIENCE. 



[Vol. VII. , No. 160 



above which the culminating peak rises sharply. 

 Each portion of the length of the ravine has its 

 characteristic features or habitus ; and, however 

 irregular these minuter details may be, they sel- 

 dom mask or obscure the characteristics of the 

 larger ones. 



Imagine, then, a great volcanic cone, on which 

 erosion has made considerable though not ex- 

 treme progress, to be truncated at about one-third 

 to one-half the height above the base, the upper 

 half or two-thirds of the altitude removed, and 

 a vast depression excavated in the remaining 

 portion. The steep wall-faces of this excavation 

 would cut the buttresses and ravines a little be- 

 low the maximum depths of the latter. The 

 crest-line at the edge of the pit, as we followed 

 around its periphery, would rise sharply to go 

 over the buttresses, and descend as sharply to cross 

 the beds of the old ravines, making it a jagged 

 edge. It is so at Crater Lake. As we ascend 

 the ravines, we find them growing deeper and 

 steeper, until at last their upper courses are sud- 

 denly cut off at the brink of the great pit. On 

 either hand rises the old buttresses many hun- 

 dreds, sometimes more than a thousand, feet 

 above us. The imagination only can picture the 

 restoration of the missing pile and the upward 

 continuation of the great ridges and furrows now 

 ending so strangely, and otherwise unaccounta- 

 bly, upon the brink of this deep gulf. Whether 

 the mountain culminated in a sharp and lofty 

 cone like Mt. Pitt and Mt. Scott to the south of it, 

 and Mt. Thielson to the north, or was a somewhat 

 natter structure like Union Peak to the east of it. 

 is more doubtful. The general configuration of 

 the ravines, and the absence of large masses of 

 tuff, or fragmental ejecta, in the original pile, indi- 

 cate the flatter, or dome-like form ; and this is de- 

 cidedly the prevailing form of mountains in the 

 Cascade Range, though many sharp peaks are 

 scattered among them. What dire catastrophe 

 has destroyed this cone ? 



Great pit-craters, or, as I have termed them 

 elsewhere, ' calderas,' are not very common. Still 

 they exist in several parts of the world ; and of 

 some of them we know the history, or may infer 

 it with considerable confidence. 



There are three or four large ones in the Ha- 

 waiian Islands. One is on the summit of Mauna 

 Loa ; a second is the famous Kilauea ; and the 

 largest ;m<l most wonderful of all is the immense 

 caldera of Haleakala, on the island of Maui. But 

 none of them are so large as Crater Lake, nor so 

 deep. The origin of these I have endeavored to 

 explain in a paper on the Hawaiian volcanoes, 

 published in the ' Fourth annual report of the 

 U.S. geological survey.' In the correctness of this 



explanation I feel great confidence. The evidence 

 of it is summed up in the paper referred to. 

 These ' craters,' or calderas as they are there called, 

 appear to have been formed gradually, through the 

 melting of the cores of the mountains by super- 

 heated lavas (i.e., lavas of higher temperature 

 than is necessary for the fusion of their materi- 

 als), rising from great depths in the earth through 

 volcanic pipes. The peculiarities of the Hawaiian 

 lavas are the absence or rarity of explosive or 

 violent action, their high temperature and great 

 liquidity. They rise in the volcanic pipes, and 

 remain stationary at a certain altitude ; and in 

 Kilauea they maintain large lakes of lava open to 

 the sky in a state of continuous fusion. But be- 

 neath the floor of the caldera they form lakes of 

 still greater extent. Eruptions occur from time 

 to time ; but the lavas, instead of overflowing 

 from the summit of the volcanic pile, burst out 

 miles away from it, and far down the gently 

 sloping sides of the cone at levels thousands of 

 feet lower than the crater. The lavas beneath 

 the caldera are drained ; and the upper portion of 

 the mountain, robbed of the liquid support which 

 has held it up, sinks in. The surface-rocks, 

 being vesicular or spongy, are light enough to 

 float on the liquid lava so long as the latter main- 

 tains its level in the stand-pipe; but, when the 

 liquid is tapped off through a lateral vent in the 

 mountain-side, the upper crust settles, as would 

 the ice in a pond when the water is drained from 

 beneath it. The evidence of this action at Ki- 

 lauea, on Mauna Loa, and still more emphatic- 

 ally on Haleakala. is very clear and unmistakable. 



But there is another class of calderas, formed 

 by a mode of !volcanic action which is in the 

 strongest possible contrast with the foregoing ; 

 and we are not left in any doubt as to its general 

 nature, for it has been witnessed and reported 

 upon by competent authority. In the islands of 

 the East Indian archipelago, stretching from the 

 Straits of Sunda eastward to the island of Timor, 

 is found a chain of volcanoes comprising hun- 

 dreds of individual cones. During the period of 

 occupation of these islands by the Dutch, numer- 

 ous eruptions have occurred ; and the most char- 

 acteristic feature of them has been their terrible 

 and devastating energy. Some of the volcanoes 

 are truncated cones, with large calderas in their 

 summits. Two of them have been formed 

 within the historic period, and accounts of their 

 formation have been preserved. One of these, 

 in the summit of the volcano Papandayang, on 

 the island of Java, was formed in 1772, by an 

 explosion rivalling in destructiveness and energy 

 the outbreak of Krakatoa in 1888. The other is 

 found in the summit of the volcano Tomboro, on 



