244 



SCIEJ^CE. 



[Vol. VIII., No. 188 



periodicity which might be due to the gi-avitational 

 action of the sun or moon, with the result that the 

 periods most likely to be discoverable are semi- 

 annual and annual, he gives a method of com- 

 bining the monthly numbers so as to eliminate any 

 shorter periods. The author then applies these 

 methocJs to Mallet's list of European earthquakes, 

 to New Zealand earthquakes (1869-70), to the East 

 Indian Archipelago (1873-81), to Chih earthquakes 

 (187a-81), and to the Grecian Archipelago (1859-73). 

 The resulting numbers are plotted in two sets of 

 curves ; the one showing the annual period, the 

 other the semi-annual period if there be one. 

 The curves, excepting that for the East Indies, 

 show clear indication of a semi-annual period, but 

 the author finds reason to doubt whether it is to 

 be attributed to the gravitational cause which sug- 

 gested the search for it. In considering tlie pos- 

 sible effect of atmospheric changes, it is suggested 

 that earthquakes, frequently at least, are not local 

 phenomena, and their causes may be sought at a 

 considerable distance from the place where they 

 occur ; as, for instance, changes of pressure over 

 the continent of Asia or over the Pacific Ocean 

 might cause variations in the strains along the 

 littoral line between them, and so might be a 

 determining cause of earthquakes in the Japanese 

 area. Pursuing this idea. Dr. Knott finds a possible 

 or probable cause of the winter earthquake maxi- 

 mum (which his annual curves show in both the 

 northern and southern regions) in the accumula- 

 tions of snow over continental areas and in the 

 annual change of barometric gradients. 



The remainder of part i. is occuj)ied by Prof. R. 

 Shida, with two papers, entitled ' Automatic cur- 

 rent recorder,' and 'On earth-cuiTents.' In the 

 former he has described and figured an instrument 

 for automatically recording the strength and di- 

 rection of a varying electric current. In the lat- 

 ter he has collected a brief account of what has 

 been done in the way of observing earth-currents, 

 adding observations of his own which seem to in- 

 dicate, that. Mobile the magnetic declination and 

 earth -currents vary in a similar way, the latter 

 changes are not the cause of the former, inas- 

 much as an increase of western declination cor- 

 responds to an increase of current flowing from 

 north to south along a telegraph-wire, not to a 

 decrease, as should be the case if the connection 

 were causal. The author also discusses the possi- 

 ble connection of earth-currents with earthquakes. 



Transactions of the Seismological society of Japan. Vol. 

 is. part 2. Yokohama, 1886. 



The second part of this vol. ix. , separately bound, 



is entirely occupied by John Milne with an account 



of the volcanoes of Japan, mainly historical and 



descriptive. The author gives a list of over forty 



works which have been consulted in its prepara- 

 tion, of which twenty-six are in Japanese, a con- 

 siderable proportion being in manuscript. The 

 information thus gathered from previous writers 

 is supplemented by extended personal observations 

 by the author in frequent journeys made for the 

 ]Durpose during his residence for a dozen years or 

 more in Japan ; and it is these accounts of his 

 own explorations that the ordinary reader will 

 find most interesting. Among them may be 

 especially mentioned his visit to Oshima(p. 78 ff.), 

 wliere he had an opiDortunity of looking down 

 into the open crater of an active volcano, which 

 was at the time belching forth masses of molten 

 lava to a height far above the point where he 

 stood. It must certainly have been, as he says, 

 'a sight of extraordinary grandeur.' A map is 

 given on which are marked 129 mountains of vol- 

 canic, origin, 23 being in the Kurile Islands. Of 

 these, 51 are active, 16 being in the Kuriles, and 

 11 in Yezo. Of the whole number, 89 are sym- 

 metrically formed cones, showing a more or less 

 close approximation to the theoretical outline de- 

 duced by Milne in the Geological magazine, and 

 by Becker in the American journal of science, and 

 again discussed by Milne in this paper. From 

 several considerations the author infers that the 

 volcanoes of the Kuriles are of more recent for- 

 mation than those of Japan. 



This part, ii., of vol. ix.,is issued from the ofiice 

 of the Japan Mail in Yokohama, instead of from 

 Tokio as heretofore, and there is also an entii'e 

 change in its outside appearance. There are 

 numerous errors, which show that the English 

 proof-readers in that otfice are not yet quite per- 

 fect. The word ' ejectamenta ' has proved espe- 

 cially puzzling to them, being misprinted in six of 

 the nine places where it occurs. 



The Japanese Transactions of this society have 

 reached vol. iii., which contains papers on ' Earth 

 tremors,' by Milne ; on the ' Earthquake of Oct. 

 15, 1884,' by Sekiya ; and on ' Air-waves and sea- 

 waves,' by Wada. 



The volume recently issued in the International 

 scientific series, on ' Earthquakes,' by John Milne, 

 is also before us ; but this article has already 

 reached such a length, that its consideration must 

 be postponed to another time. 



C. G, EoCKWOOD, Jr. 



From Nature we learn that Japan has thirty- 

 seven periodicals devoted to education ; seven of 

 these are medical, with a monthly circulation of 

 13,514 ; nine treat of sanitary matters, and two 

 of pharmacy ; twenty-nine are what might be 

 termed popular scientific journals, and have a 

 circulation of 70,000. 



