June 1, 1883.] 



SCIENCE. 



485 



Madagascar and New Zealand are of older separa- 

 tion, the latter approaching the restinseln. 



The term 'oceanic' is discarded for the second 

 group, because islands maybe formed denovo close to 

 continental shores ; but the term proposed ( ' original ' ) 

 is not altogether satisfactory, as it does not express 

 precisely what is meant. The iirst subdivision (vol- 

 canic islands) contains the most important examples, 

 which have sometimes, from tlieir considerable age 

 and altitude, acquired peculiar and local organic 

 forms. The second subdivision (heaped-up islands) 

 includes those of coral and of sand, on which the dry 

 surface is due to wave and wind action. These are all 

 low and monotonous. The third subdivision includes 

 portions of the sea-bottom laid bare by non-volcanic 

 action, either by local elevation "or by withdrawal 

 of the sea formerly held at a higher level by the local 

 attraction of mountains or ice masses that have now 

 disappeared." A single example of recent formation 

 is given, — the so-called ' Gulf-stream island,' north- 

 west of Novaya Zemlya, where the Dutch navigators 

 of 1594 found a sand-bank in seventeen fathoms of 

 Water. Peschel's error of placing the Japanese and 

 Philippine islands among the volcanic is corrected: 

 they are included among the continental, as both con- 

 tain a series of old nou- volcanic rocks. 



W. M. Davis. 



LETTERS TO THE EDITOR. 



A nevT form of battery-cell. 



In the ordinary voltaic element, two solid plates 

 are acted upon unequally by one or more liquids. 

 About three years ago, it occurred to me to construct 

 a battery-cell with three non-miscible liquid strata, 

 and no solid plates ; which I did, as follows : in a 

 small beaker-glass I placed successively layers of mer- 

 cury, dilute sulpliuric acid, and a solution of iodine 

 in ether. Upon connecting the uppermost and 

 lowest layers with insulated wires, and introducing a 

 coarse galvanometer into the circuit, I obtained evi- 

 dence of a fairly strong current of electricity. Hav- 

 ing neither time nor opportunity to jjursue the matter 

 further, I put it on record now in order that any 

 student who happens to be interested in the subject 

 may carry out the investigation. Theoretically, a 

 three-liquid cell is interesting, because its internal 

 resistance ought to diminish with rise of tempera- 

 ture. In this respect it might be very different from 

 the usual voltaic elements. Possibly a combination 

 of solid plates with the upper and lower liquids 

 might give a cell having an internal resistance con- 

 stant for varying temperatures. F. W. Clakke. 



Correcting compass deviation. 



Some years ago, frequently recurrent shipwrecks 

 from magnetic disturbance in the Gulf of St. Law- 

 rence directed my attention to the subject of improv- 

 ing the mariner's compass, or supplementing it in 

 some way which would make its Indications trust- 

 worthy. The causes of the shipwrecks which I have 

 mentioned seemed to be deijosits of iron ore near 

 the shore, so extensive in their area as to render the 

 compass-reading false and misleading. The j^roblem 

 of improving the compass is an important one ; for, 

 apart from such risks as those which beset naviga- 

 tion in the Gulf of St. Lawrence, the deviation on 

 board ship due to the presence of iron in tlie struc- 

 ture or cargo of the vessel is an element of some 

 uncertainty, and danger even, when all the devices 

 known to the mariner's art are used to correct the 

 readings. 



My first attempt was to so dispose a series of small 

 flat magnets, fastened across a strip of aluminum, 

 that the strip as such, when poised at its centre, 

 pointed east and west. 



Poised concentrically with the strip at such a dis- 

 tance as to avoid mutual influence, I placed a light 

 magnetic needle of a length equal to that of the 

 strip. When strip and needle were near enough to a 

 jjiece of iron to be attracted by it, one of the two 

 acute angles formed by them indicated the position 

 of the disturbing iron ; and this inclination at an 

 acute angle promised to be of value in solving the 

 problem of correcting compass-readings. But mag- 

 netic influence on the large scale which jirevails on 

 shipboard proceeds from distant centres along large 

 curves, in which terrestrial and local forces merge, 

 which are not attractive, but simply directive ; so 

 that when I tried my device on a steamer under very 

 favorable experimental circumstances, as the mag- 

 nets, large and small, were directed into cuiwes so 

 great as to be practically straight lines, the needle 

 and strip were always at right angles with each other. 

 Were it feasible to use a very long magnetic strip at 

 sea, my device might be available ; but, so long as 

 ships roll and pitcli on the ocean's unruly surface, 

 the dimensions of the ordinary compass must remain 

 as they are. Since abandoning the fragile little 

 model which I launched with some expectations 

 long ago, I have frequently reverted to the problem 

 it was intended to solve ; and it has occurred to me, 

 that were an electro-magnet poised so as to be in 

 constant and free communication with a battery, and 

 were it possible to make it, when desired, so intense 

 in its power that its induction affecting the iron 

 of ship or cargo should increase the deviation which, 

 when less intensely excited, would affect it, then the 

 direction of the deviation would be, of course, known 

 by the direction of the increase of that deviation, 

 and the problem of correcting the compass-reading 

 would be advanced a step. The intensity of the 

 electro-magnet would yield such results as a long 

 (impracticably long) magnetic strip. The electro- 

 magnet would require to be so constructed as to be 

 capable of developing the utmost intensity possible; 

 and the current sent through it should be control- 

 lable at will, so that the indications at moderate and 

 highest intensity might be compared. I have neither 

 tlie skill nor opportunity to carry out the suggestion 

 here given, and publish it in the hope that some 

 competent man of science may be able to embody it 

 in a practical and useful form, Geokge Iles. 



Montreal, May 25, 18S3. 



MAINE'S EARLY LAW AND CUSTOM. 



Dissertations on early law and custom. By Sir 

 Henry Sumner Maine, K.C.S.I., LL.D., 

 F.R.S. New York, Henry Holt §■ Co., 1882. 

 402 p. 8°. 



When a new book bj- Sir Henrj' Maine is 

 announced, we expect to have something to 

 read worth reading. Nor have we e-\'er been 

 disappointed. The author of ' Ancient law ' 

 has alwaj's something interesting, snggestive, 



