220 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. [DeC. 1 9, 



was formed in 1 707, it went on increasing irregularly, and sometimes 

 was lowered on one side while it gained height on the other. At 

 diiferent periods also during the growth of the island, isolated rocks 

 rose up in the sea at diiferent distances from its shore, some of them 

 appearing and disappearing at intervals. There were many evidences 

 of eruption before a visible crater was at length formed, so that we 

 may infer that an intumescent mass of pasty or fluid trachyte was 

 forcing up the top of the hill, as we see lava-currents, when they meet 

 with an obstruction, swell up because they are encased by and con- 

 fined within a solid exterior, the sides of which often slope at an angle 

 of more than 40°. The large open rents seen on the surface of the 

 Old Kaimeni or Hira attest the distension of that island, during the 

 injection of lava beneath it. In a word, the whole history of these 

 central islands shows that they owe their origin to the successive and 

 intermittent action so characteristic of volcanos, and lends no support 

 to the hypothesis of a single paroxysmal explosion, by which either a 

 gigantic mountain or crater can be formed at one effort. Had the 

 denuding action of the sea never removed the central portions of the 

 ancient cone, all those masses of brown trachytic lava and pumice 

 which have now gone to the production of the central volcano, called 

 the Kaimenis, would have been expended partly in the filling of fis- 

 sures with melted matter, forced upwards, partly in the outpouring 

 of lava, and ejection of scoriae from a permanent central vent. For 

 in some cases, as in the Sandwich Islands, we see craters much loftier 

 than that which crowns Etna emit streams of lava of enormous vo- 

 lume. But it happens more commonly in volcanos, that, as they gain 

 in height, the pressure of the central column of lava overcomes the 

 resistance offered by the sides of the cone, so that the latter give way 

 at some points. There can therefore be little doubt that a large pro- 

 portion of the materials now composing the Kaimenis would, if the 

 great dome had remained entire, have been emitted in the form of 

 lateral cones. Had this occurred, the volcanic strata now encircling 

 the Gulf of Santorin would have been intersected by veins and dikes, 

 whereas none of the geologists who have visited Santorin make any 

 allusion to such dikes, and Mr. Edward Forbes tells me he observed 

 none of them in any of the three outer islands, Thera, Therasia, and 

 xlspronisi. We must consider therefore these three masses as the 

 basal remains of a large dome or cone, so far removed from the ori- 

 ginal centre of eruption as not to have been subject to injection from 

 below. 



As the theory of denudation requires us to suppose in the case of 

 Santorin an oscillation of level, that is to say, first the gradual rise of 

 a cone of submarine origin and secondly its partial submergence, it is 

 worthy of remark that Lieut. Leycester states that on the east side of 

 Thera there is a road now twelve fathoms under water, which formerly 

 led from Perissa to Camari, and which was above water before the 

 earthquakes of 1650, in which year a volcanic eruption occurred in 

 the sea about three miles and a half north-east from Cape Colombo 

 in Thera, where vapour and flames were thrown out and the sea was 

 covered with pumice, and where after some months a shoal (fig. 4, A, 



