which have been referred to above; the eruptions of Kilauea are of a quiescent type, in which many of these reactions have taken place before the gases escape, and observations show that the percentage of water vapour is less in those collected most directly from the lava as it ascends from the central pipe, dropping to only 4 % in the col- lection made most directly from the issuing lava. In more violent eruptions, where the gases issue more nearly in the composition with which they are released from the magma, the proportion may well be less than this, and even drop to nothing. Brun's observation that the fumaroles nearest the crater of Vesuvius, having tempera- tures of 400 C. or more, yield very small amounts of water vapour, has been confirmed by Ferret on Vesuvius and Ponte on Etna, but the most pertinent fact is, probably, the presence of chlorides of iron, magnesium and aluminium in the ashes of the cone, and within the crater, after violent eruptions of Vesuvius. Being deliquescent these minerals could not have crystallized in the presence of satu- rated steam, and equally could not have been formed in the presence of superheated steam, as they would then have been decomposed with the formation of oxides, at temperatures even below 500 C., much lower than that of the escaping gases of the eruption. There is, therefore, reason to suppose that water vapour, as such, may be absent from, or form a trivial proportion of, the gases issuing from the vent of a volcano in violent eruption ; it would be formed, after issue, as the result of reactions between the components of the highly heated ascending column of gases, and by union of hydrogen with the oxygen of the atmosphere. This supposition is, moreover, in keeping with the modern trend of conjecture regarding the nature of the fluidity of the magma, according to which the effect formerly ascribed to water is now frequently attributed to hydrogen, either free or in combination with other elements than oxygen.
Earthquakes (see 8.817, 11.659). Not much has been added in more recent years to our knowledge of the character and effects of earthquakes. The ultimate cause is still very incompletely understood, but it has become clear that faults, of structural geology, are much less directly connected with earthquakes than was formerly supposed. It still remains established that the vast majority of earthquakes are caused by the shock resulting from dislocation of the solid material of the earth's crust, due to fracture consequent on a state of strain, but the detailed examination of the Californian earthquake showed that this strain was radically different in character, and the result of some quite different cause, from that which produced the San Andreas fault, along which visible displacements took place. The earth- quake was not, in fact, an incident in the growth of the fault, nor the fault the cause of the earthquake; the real connexion of the two being that the fault produced a plane of weakness, along which relief of the growing strain took place, and so controlled the local distribution of the intensity of disturbances. It is probable that detailed investigation would yield a similar result in the case of other earthquakes, where a visible connexion be- tween the violence of the earthquake and the existence of struc- tural faults has been observed, for another result of the examina- tion of the Californian earthquake was to show that the fractur- ing of the crust, to which the felt earthquake was due, was but the secondary result of a more deep-seated disturbance or bathyseism, and that this bathyseism, the exact character and depth of which are still unknown, gave rise to the long-distance records commonly attributed to the earthquake proper (see SEISMOLOGY). Though, as has been said, nothing is known of the character of the bathyseism, it may be connected in some way with the changes of volume consequent on a redistribution of the chemical elements composing the material in the interior of the earth, which are referred to in the section dealing with the origin of mountain ranges; according to the ingenious specula- tions of Mr. W. H. Goodchild it is probable that some of these changes may take place with great, even explosive, rapidity, and so might easily give rise directly or indirectly to both the earthquake proper and the long-distance record.
The study of the periodicity of earthquakes has giyen some curious and interesting results. According to the tabulation of 20 years' record of the Italian seismological service there is, in Italy, a well- marked diurnal periodicity, giving a well-marked maximum fre- quency at about two hours after midnight, and a period of low fre- quency extending from about 9 A.M. to 5 P.M. with a minimum at about two hours after midday, the maximum frequency being almost twice as great as the minimum and the number of earthquakes occur- ring during the 12 hours of the night being very close to half as many again as the number recorded during the 12 hours of the day. The subject has been elaborately investigated and the periodicity shown to be a real one, not to be accounted for by any imperfection of the
record. Though certainly real, the particular form of periodicity applies only to Italy, other regions differing in this respect, and no satisfactory explanation has been offered. It is evidently not due to tidal stresses set up by the attraction of the sun, for the periodicity completely disappears when the record is tabulated by lunar instead of solar time; there is, however, another peculiarity of frequency which does seem attributable to this cause. In any record of suffi- cient extent to be usable, it has been found that the relative propor- tion of day shocks to night shocks is slightly greater during summer, and slightly smaller during winter, than the general average for the whole year. The variation from the mean is small, not exceed- ing 6% in the most extreme case which has been investigated, and is generally less; it can only be detected when a large number of records are investigated, but as the same relation has been found repeated in the only two cases where the record has been discussed in terms of lunar hours and declination, it may be accepted as real, and very probably connected with the different distribution of the tide-producing stresses during the day and night, according to the varying declination of the sun and moon. Whether this conclusion be accepted or not, the very smallness of the effect shows that earthquakes are uninfluenced, except possibly, and very insignifi- cantly, as to the actual time of occurrence, by any cause exterior to the earth; they are a purely terrestrial phenomenon, due to actions which take place within the outer crust of the earth and, therefore, the study of them is distinctly a branch of geology.
Origin of Mountains (see 11.659). The problem of the origin of mountain ranges, and incidentally of the major inequalities of the surface of the earth, stands in a very different position from that which it held in 1910. At that time it was recognized that in some of the elevated tracts of the earth there was no obvious connexion between geological structure and orography, but in other cases, of which the Alps were a type, it had been observed that there was very intense disturbance and compres- sion of the rocks, that the strike of the folds, into which the rocks had been thrown, agreed with the general direction of the range; and in these cases, regarded as" true " mountain ranges, it was believed that the surface elevation was due to a thickening of the crust consequent on the compression indicated by the disturbance of the rocks (see Plate).
Even in the Alps there were difficulties: the compression is not simply towards the principal crest of the range, but large blocks of strata had been thrust from the southern limit of the range to near the northern, and though it was possible to explain this by a gradual southward migration of the central axis of compression many diffi- culties arise in the course of such explanation. Another point, which could not easily be met, is the want of a relation between the degree of compression and the height of the mountains. The Himalayas, also regarded as a true mountain range of Alpine type, are about twice as high as the Alps, but the amount of compression of the rocks is, if anything, distinctly less. In the course of the decade 1910-20 other ranges were studied ; the Andes are now known to exhibit much less compression of the strata than the Alps, less in fact than many lowland regions which give no indication of having ever formed lofty mountains; but the most striking case of want of accord between the geological structure of rocks and the relief of the surface is in the region of the Pamirs. Here, between the plains of Russian Turkestan, on the W., and of Chinese Turkestan, on the E., and between the upper waters of the Jaxartes and Oxus, on the N. and S., is a great mountain mass, in which the general course of the ranges and river valleys, as of the trend lines of the geological struc- ture, is between S.W.-N.E. and W.S.W.-E.N.E. From the low hills on the west the general level of the surface rises to the high plateau of the Pamirs, and eastward of these comes a well-defined mountain range, the Mustagh Ata range, with peaks ranging to over 25,000 ft., running nearly N. to S. along the edge of the plains of Chinese Turkestan. So well marked and definite is this range that the late Prof. Suess, in his great work on The Face of the Earth, concluded that it must be structural, and that the trend-lines of structure as seen to the westward must bend round to follow the course of the mountain ranges. The earlier observations of Dr. F. Stoliczka, and the recent explorations of Sir Henry Hayden, have alike shown that this is not so, and that the general strike on the Pamirs is continued across the range to the plains of Chinese Turkestan.
In both of these mountain regions, the Andes and the Pamirs, and especially in the latter, large parts of the surface show the rounded outlines and moderate slopes of a lowland topography which has been subject to a prolonged period of subaerial denudation, the mountainous character, where present, being due to quite recent erosion by streams and rivers. In other words, the indications in both regions are that the present elevation is due to a simple uplift, quite independent of, and of later date than, the compression which is evidenced by the surface rocks. In parts of the central Himalayas there is evidence leading to a similar conclusion, and most of the difficulties in reconciling the structure with the larger features of the surface relief of the Alps would disappear if it were accepted.
Meanwhile, an entirely different line of research had led to a simi-
