July 2, 1903] 



NATURE 



199 



Tables of Four-figure Logarithms. 



For many scientific computations it is sufficiently accurate 

 to work to four figures, but there have been complaints that 

 the usual tables of four-figure logarithms are not accurate 

 in the fourth figure. Thus, log i 019 is given as 00080, 

 whereas it ought to be 00082. The errors are met with 

 only in numbers from 1000 to 2000. In consequence of this, 

 some such tables are accompanied by a table specially in- 

 tended for numbers between 1000 and 2000. Many physicists 

 and chemists refuse to use four-figure tables for this reason, 

 and advocate the use of five-figure tables, in spite of their 

 greater size and the waste of time. 



I beg to point out that Mr. J. Harrison has got over the 

 difficulty in a very simple manner in the four-figure table 

 published by him recently in his book, " Practical Plane 

 and Solid Geometry." Even he, however, cannot avoid a 

 possible error of i in the last figure. The first ten rows of 

 differences have been replaced by twenty. The rest of the 

 table is unaltered. I give a specimen of an old row of 

 figures and how it is replaced. The cause of inaccuracy in 

 the old system is apparent at once. 



and, hovering there for an instant, they cooled and con- 

 tracted, and sank slowly down through the water. When 

 the bubbles are formed in rapid succession, the phenomenon 

 is one of great beauty, as their surfaces are extremely 

 brilliant, being formed of mercury freshly drawn out before 

 rising into the water. The mercury used in this experi- 

 ment was the ordinary commercial article, and not freshly 

 distilled. Grease had, however, been removed from it by 

 boiling with a solution of caustic potash. Tap water was 

 used. 



I have since found that these mercury bubbles are easily 

 produced in a variety of ways. The most striking form 

 of the experiment is perhaps as follows : — About 30c. c. of 

 mercury are poured into an evaporating dish and covered 

 by a depth of about i-7cm. of water. Bubbles are now 

 formed in the mercury by forcing air under its surface 

 through a bent glass tube drawn to a fine nozzle. When 

 th" bubbles reach a certain size they become pyriform and 

 draw out from the surface of the mercury, and, rising 

 through the water, float on its surface. The bubbles so 

 formed have considerable stabilitv, and usuallv last for 



It is to be hoped that all four-figure logarithm tables 

 will in future be printed in this way. The Board of 

 Education is now printing tables of this kind for use in 

 evening science classes. John Perry. 



Ship's Magnetism. 



In a review of my book on the subject of the " Deviations 

 of the Compass in Iron Ships " which appeared in Nature 

 of June 18 and in the last paragraph, there are state- 

 ments to which I would take exception. In this para- 

 graph the reviewer finds " food for reflection " in that, after 

 defining C.G.S. units in the introduction of my book, I 

 stick to inches and to other units in the text and charts. 



In view of the fact that every measurement in a ship is 

 recorded in feet and inches, whether by constructors, 

 engineers, gunners, or navigators, to have introduced the 

 centimetre as a common unit of measurement in the text 

 and tables would have been a serious drawback to the 

 utility of the book. 



Again, the values on the charts of horizontal and vertical 

 force are given in terms which have been found most con- 

 venient in the several computations, whilst not detracting 

 from their value as exponents of the changes of terrestrial 

 magnetism which a ship may encounter during a voyage. 



Whilst introducing the student to the modern C.G.S. 

 units, the use of the British units is too recent for neglect- 

 ing to mention them, hence their retention on the map at 

 p. 16, accompanied by the necessary multiplier for convert- 

 ing them to C.G.S. units if required. 



A table of errata has been published for some weeks, and 

 sent to all known recipients of the book. 



June 20. E. W. Creak. 



Mercury Bubbles. 

 Recently during the course of an experiment I had 

 occasion to boil water in presence of mercury. After 

 ebullition had been going on for some time, I noticed that 

 occasionally steam forming below the surface of the 

 mercury carried with it a pellicle of mercury as it rose 

 through the water in the form of a bubble. When it 

 reached the surface of the water the pellicle usually burst, 

 and the mercury fell back as a drop. By adjusting the 

 intensity of ebullition, it was possible to bring the two 

 liquids into such a state that, comparatively frequently — 

 say ten times per minute — steam bubbles covered with 

 mercury rose through the water and floated on its surface, 



NO. 1757, VOL. 68] 



I I5~30 seconds before bursting. One having a diameter of 



i about I -gem. was timed to have lasted for 756 seconds, 



j floating on the water. When the break does occur, it has 



j explosive violence, and drops of mercury are thrown through 



I a considerable distance. The bubbles which reach the 



! surface of the water intact do not vary, as a rule, much in 



size, the maximum diameter observed being 20cm. and the 



minimum 1 8cm. The weight of mercury in the bubbles 



may be determined by tioating them into a watch glass, 



and weighing the mercury which falls down from them on 



bursting. There is always more than 0150 gram in the 



I pellicle, and rarely more than o 200 gram. The mean of 



i ten weighings gave o 177 gram as the weight of mercury 



in these bubbles. From these data it appears that the 



\ mean thickness of the mercurial pellicle is oooicm. At 



the thinnest part, however, it must be much thinner, for, 



as the profile view shows, each bubble carries a drop of 



I varying dimensions hanging from its lower pole. The 



j bubbles float immersed nearly to their equator. In the 



\ majority of cases they remain covered with a skin of water, 



so that the meniscus of the water is not depressed round 



the floating bubble, but is raised round it. The skin of 



water may be made to retreat from the upper pole, or to 



aggregate itself into droplets on the surface of the bubble, 



without causing the rupture of the bubble, by the addition 



i of a small amount of spirit to the water. The complete 



; absence of the water skin from the mercury pellicle may be 



demonstrated by the dulling of the surface of the latter 



1 when breathed upon. 



The high surface tension of the water does not seem 

 necessary to the phenomenon. Mercury bubbles in every 

 way similar to those just described may be formed under 

 methylated spirit, and will float upon its surface ; also the 

 addition of a slight contamination to the water, such as 

 oil or soap or spirit, does not make the mercury film of 

 the bubble completely unstable. But when large quantities 

 j of these impurities are added, the bubbles seldom last more 

 than a moment on the surface of the water, although even 

 in the presence of these impurities they may last as long 

 as 25 seconds. 



The depth of the overlying water is of importance in 

 I the ease of producing stable bubbles by this method. If 

 i it is deeper than 2cm. the bubbles usually break before 

 I getting to the surface ; this is probably due to the change 

 of pressure during the rise of the bubble through the 

 water, and consequent excessive expansion causing rupture. 

 I If the water is less than i-5cm. deep the bubble formed 

 i swells up to a great size (2-3cm. in diameter) before it 



