I04 



NATURE 



[June 2, 1898 



"At the same time the rain fell after sixty-six days' drought, no 

 such instance of dry weather in the West Indies was remem- 

 bered). The next minima (of wet) would correspond to i862'5 

 (great earthquake in Greece, 26 Dec./6i, and eruptions of 



Vesuvius accompanied by earthquakes), and the last, to 1897 "5, 

 which fairly corresponds to the great earthquake of Assam, 

 so fully noticed in your journal, as one of the most intense of 

 modern times. Moreover, these figuies may be 

 presented otherwise. Taking the great earth- 

 quake of Lisbon as of date 1755, roughly we have 

 the annexed succession of years showing at the 

 two extremes the dates (approx. ) of two of the 

 greatest earthquakes of modern times, and to 

 some extent showing that thirty-five years re- 

 presents a period of maximum earthquake action, 

 and agreeing roughly with the intervals of extreme 

 drought and with periods of great volcanic activity. 

 As regards the year 1825, it is interesting to 

 note that Mallet's Catalogue gives for July 26 

 and 27 of that year, "One of the most tremendous 

 hurricanes on record occurred in the West Indies." 

 Of course a great deal has to be said as to the locality of the 

 earthquakes, and as to the volcanoes to be considered. I 

 certainly look on those of the Andes Cordillera as of prime 

 importance by their influence on the upper currents. 



Royal College of Science, Dublin, J. P. O'Reilly. 



May 21. 



Ebbing and Flowing Wells. 



A CASE somewhat resembling those previously described 

 (Nature, May 12, p. 45, and May 19, p. 52), occurs on the 

 dormant volcano of Barren Island in the Andaman Sea. The 

 only (comparatively) fresh water to be found on the island 

 reaches the surface in the form of hot springs, which gush out 

 close to the shore at the breach through the ancient cone. The 

 springs are due to the percolation of the drainage water beneath 

 the most recent lava streams, which have not yet fully cooled 

 down. The level of the springs rises and falls with the tide, 

 and the lower part of a well, which I caused to be dug. in the 

 ash about twenty yards from the shore, filled with hot water at 

 the flow of the tide, and ran dry at the ebb. The bottom of the 

 well was between tide levels. The water is brackish, but rather 

 less so at high than at low tide, the reason of which appears to 

 be as follows. The porous volcanic materials of the island below 

 sea level are saturated by the water of the sea, the surface of this 

 inland subterranean water rising and falling in connection with 

 the rise and fall of the sea tide The drainage of the amphi- 

 theatre, then, soaks downwards until it reaches the inland 

 salt water, over which, on account of the difference in specific 

 gravity, it flows onward to the sea. At high tide, therefore, the 

 drainage reaches the sea through materials which have been 

 comparatively little wetted by salt water, while at low tide it 

 percolates through, and washes, ejecta from which the salt water 

 has just retired. The phenomenon is, of course, complicated by 

 the difference in time between the inland tide and that at sea. 



The springs are described in some detail in Memoirs Geol. 

 Surv. Ind., vol. xxi. p. 274 (also Records G.S.I., vol. xxviii. 

 pp. 31, 34). F. R. Mallet. 



May 25. 



NAVIGATION. 

 'VT.'WIG.^TION, in its widest sense, is generally de- 

 -'■^ fined as the art of conducting a ship from port to 

 port, and may conveniently be divided into coasting and 

 guiding the path of a vessel across the trackless ocean. 



Coasting is principally pilotage, assisted by a few rules 

 based on geometry and plane trigonometry, combined 

 with a knowledge of that oldest and most valuable of 

 seamen's friends, the mariner's compass. A knowledge 



NO. 1492, VOL. 58] 



of the compass in Europe is much older than is generally 

 supposed. It was certainly used as far back as the 

 beginning of the thirteenth century. 



The compass plays a still more important part in 

 deep sea navigation (with which this paper is more 

 particularly concerned), which is so closely allied to 

 nautical astronomy that in one sense of the word it 

 includes it, whilst in another it distinguishes the terres- 

 trial methods of finding the position of a, ship at sea, 

 from the more accurate methods of locating her where- 

 abouts, that the researches and labours of the astronomer 

 have placed at the disposal of the navigator. 



The earliest efforts of the seaman, when he ventured 

 out of sight of land, were directed by the compass, which 

 of late years has been immeasurably improved, and by a 

 log for measuring the rate of sailing, which has become 

 almost as obsolete as the plane sailing and the plane 

 chart by which he estimated his position. This method, 

 proceeding on the assumption that the earth's surface is 

 a plane, was fairly accurate for moderate distances near 

 the equator, or even in higher latitudes if the vessel sailed 

 on, or near a meridian, but was quite incapable of 

 measuring differences of longitude, and if used, for 

 instance, on a westerly course from Cape Clear, would 

 produce an enormous error, if the departure or westing 

 was taken as the difference of longitude. Owing to the 

 uncertainty and variability of the wind, sailing vessels 

 altered their course so often that, to save the labour of work- 

 ing out the difference of latitude and departure for each 

 course and distance by trigonometry, the traverse table 

 was introduced. It is sitnply the tabulated values of the 

 sides of a number of right-angled triangles, where the 

 hypothenuse is the distance, the perpendicular the de- 

 parture, the base the difference of latitude, and the course 

 the given angle. By means of this table it was easy to 

 get the difiference of latitude made good, by taking the 

 difference between the sum of the northings and south- 

 ings, and the departure made good, by subtracting the 

 eastings from the westings, or vice versa. This was 

 called resolving a traverse. The inability of plane 

 sailing to afford the difference of longitude led to the 

 introduction of parallel sailing, middle latitude sailing, 

 and Mercator's sailing, and the inestimable chart that 

 bears the name of the latter. It is easily demonstrated 

 by solid geometry, that the arc of a parallel of latitude 

 between any two meridians is equal to the corresponding 

 arc of the equator multiplied by the cosine of the latitude ; 

 so that if a ship sails on a parallel, it is a simple 

 operation to convert her meridian distance or departure 

 into difference of longitude. But a ship does not always 

 keep to a parallel ; in sailing, however, from point to 

 point, she must leave one parallel and arrive at another. 

 Now let the portion of the rhumb line between these two 

 parallels be conceived to be divided into infinitely small 

 parts, which will be sensibly straight hnes on each of 

 which is a triangle representing the corresponding differ- 

 ence of latitude and meridian distance. Then the de- 

 parture will be the sum of all these meridian distances, 

 and must be equal to the arc of a parallel somewhere 

 between the two extreme ones. In middle latitude sailing 

 it is assumed to be equal to the arc of the parallel that 

 lies midway between the one left and that arrived at, and 

 the difiference of longitude is obtained as in parallel 

 sailing, substituting the middle latitude for the parallel. 



Though the above assumption is not strictly accurate 

 (the real parallel always lying on the polar side of the 

 middle latitude), the results deduced from it in favour- 

 able cases are such very close approximations as to be 

 preferable to those obtained by Mercator's sailing, which 

 is theoretically irreproachable. 



About the middle of the sixteenth century, Gerard 

 Mercator introduced the chart which has since borne his 

 name, in which the meridians are all parallel and the 

 degrees of latitude increased towards the poles, and on 



