224 



NATURE 



{July 5, 1888 



The position of the rock is in latitude 14 22' 8" S., 

 longitude 42 41' 32" E., 18 miles from the island of 

 Jebel Zukur, and the same from the eastern shore of the 

 sea, and out of sight of land except in clear weather, when 

 Jebel Zukur is visible. The dangerous portion of the 

 rock is only about 40 yards in diameter, but the sound- 

 ings round for about 100 yards give indications of its 

 presence. 



Its slope is not so very steep as in some other instances 

 of coral banks in this sea. Assuming that coral after it 

 attains within a certain distance of the surface grows 

 mainly outwards, and that the almost perpendicular sides 

 of some of the Red Sea reefs are mainly the result of such 

 outward growth, the comparatively gentle slope of the 

 Avocet rock may be taken to show that it is in an early 

 stage of its development ; a view which its small size also 

 supports. 



The rock lies on the bank of soundings on the eastern 

 side of the deep-water gully up the centre of the Red Sea, 

 near its edge, and close to the point where it comes to an 

 end. It has frequently been noticed that coral patches 

 most readily form on the edges of such steep submarine 

 slopes — witness other parts of the Red Sea itself— but 

 they generally take the form of a scattered line along such 

 an edge, and it is not usual for one small and isolated 

 patch to alone make its appearance. 



This rock is nearly midway between the St. Oswald's 

 position for the Avocet and the telegraphed position 

 of the Teddington, and is about 350 yards from where 

 the Sylvia was at one time anchored. It lies about 5^ 

 miles off the direct line between the Abu Ail channel and 

 a point 3 miles west of the Zebayir Islands— the course 

 generally taken by ships. 



Seeing that transverse currents are by no means rare 

 in the Red Sea, and also that many vessels — especially 

 when bound north at night — habitually pass outside Abu 

 Ail, it is a cause for marvel that no ship has ever struck 

 this small danger before. One of the telegraph cables 

 passes close to it — so close that it is doubtful on which 

 side it lies, and the ship laying it may therefore be con- 

 sidered to have had a narrow escape. On the very 

 morning of the Avocefs loss, a large troopship passed 

 east of that vessel an hour before she struck. Evidence 

 is already forthcoming of many ships having been swept 

 to the eastward at different times, so that they must have 

 passed very close to the Avocet rock. 



The absence of a marked break on the rock is another 

 somewhat curious fact, and shows how a short heavy sea 

 without the accompaniment of an ocean swell can pass 

 over as little water as 15 feet without showing more than 

 the white horses which crown every wave when the wind 

 is strong. 



MAGNETIC STRAINS. 

 TT has long been known that when an iron rod is 

 magnetized its length is in general slightly increased. 

 This phenomenon was first studied by Joule about the 

 year 1847, and most of his experimental results have been 

 confirmed by other physicists, among whom may be 

 mentioned the names of Tyndall, Mayer, and Barrett. 



Joule enunciated the law that the elongation of a 

 magnetized rod is proportional to the square of its 

 magnetization, a law which seems to have been pretty 

 clearly supported by his experiments so far as they 

 went. Now, when iron is subjected to the action of con- 

 tinually increasing magnetizing force, a point is at length 

 reached when further increase of the force produces com- 

 paratively little effect upon the magnetization. The iron 

 is then, in popular language, said to be " saturated," and 

 is (or until lately was) commonly supposed to have 

 attained a condition of magnetic constancy, so that none 

 of the properties of the metal connected in any way with 



its magnetism would be materially affected by any increase 

 of magnetizing force, however great, beyond what was 

 necessary to produce saturation. 



Joule carried many of his observations up to the so- 

 called "saturation point," and then, perhaps naturally, 

 seems to have assumed that nothing would be gained by 

 going any further, and accordingly discontinued his 

 experiments. It is, however, a somewhat remarkable 

 fact that although his interesting discovery was soon 

 widely known, an account of it appearing in almost every 

 text-book dealing with electricity, while an exhibition of 

 the phenomenon in question became a familiar lecture 

 illustration, yet for the thirty- seven years following the 

 publication of Joule's paper it seems never to have 

 occurred to any experimenter to try what would be the 

 effect of subjecting an iron rod to stronger magnetizing 

 forces than those applied by Joule himself. Perhaps I 

 may be pardoned if I refer to the accidental circumstance 

 which led me to do so. 



In 1884, a reprint of Joule's scientific papers was issued 

 by the Physical Society, and I then read, for the first 

 time, his original memoir on the effects of magnetism 

 upon the dimensions of iron and steel bars. I had re- 

 cently been engaged in an investigation of the heat- 

 expansion of sulphur, changes in the length of rods of 

 that substance being indicated by their action upon a 

 small movable mirror which reflected the focussed image 

 of a wire upon a distant scale ; and it struck me that a 

 similar method would be well adapted for the exhibition 

 of magnetic expansions. Wishing to have the satisfaction 

 of witnessing some of these effects, I put together a rough 

 apparatus, in which the mirror principle was applied. 

 The battery employed consisted of five large bichromate 

 cells, the zinc plates of which were immersed in the solu- 

 tion by the action of a treadle, and withdrawn by an 

 opposing spring when the pressure on the treadle was 

 removed. The circuit included the magnetizing coil, a 

 galvanometer, and a contact-key. 



The first results of experiments made with this appar- 

 atus were disappointing. Everything appeared to be 

 quite right : the mirror worked perfectly, as was shown 

 by its deflection when the temperature of the iron rod 

 was slightly varied ; the iron was well annealed, and 

 there could be no doubt that the magnetizing force used 

 was more than sufficient to " saturate " it (in the popular 

 sense). Yet the elongation indicated when the circuit 

 was closed was only a small fraction of what had been 

 expected, the movement of the focussed index upon the 

 scale being, indeed, scarcely perceptible. 



The arrangement was varied in several details, and 

 further attempts were made, but without any better 

 success. In these perplexing circumstances I happened 

 to remove my foot from the battery treadle while the 

 contact key was still_depressed, and at the moment of 

 doing so I noticed a curious " waggle " of the focussed 

 image. A movement of the same kind was found upon 

 trial to occur if the zincs were lowered into the liquid 

 while the key was down. The operation was then per- 

 formed very slowly, and the exact nature of the waggle 

 became clearly revealed. As soon as the zinc plates 

 touched the surface of the liquid the index immediately 

 jumped into a position indicating a certain small elonga- 

 tion of the magnetized rod. As the zincs went in deeper, 

 this elongation at first steadily increased, but only up to 

 a certain point, after which it was diminished ; and when 

 they were completely immersed in the liquid, the focussed 

 index had returned nearly to the zero position, showing 

 that the elongation had almost entirely disappeared. 

 When the zincs were again slowly raised, the same cycle 

 of changes occurred in inverse order. 



The conclusion obviously suggested by these observa- 

 tions was one that could not be readily accepted. It 

 appeared as if the magnetizing force which had been used 

 in the first instance was too great to produce Joule's 



