270 



NA TURE 



{Jan. 21, 1886 



came close to the house and into the fyard where the poultiy 

 feed. I have not seen it for some months, nor heard of it in any 

 other garden. A. S. Mathevvs 



Edgbaston 



Curious Phenomenon in Cephalonia 



Mr. Ledger (p. 246) need not have had any doubt about the 

 correctness of the information sent him by his friend about 

 currents running from the sea into Cephalonia. It is a well- 

 known fact, and the following account of it is from Dr. John 

 Davy's "Ionian Islands," published in 1842, vol. i. 164: — 



"The next phenomenon I have to mention is very e,\tra- 

 ordinary, and apparently contrary to the order of nature : it is 

 the flowing of the water of the sea into the land in currents or 

 rivulets which descend and are lost in the bowels of the earth. 

 This occurs in Cephalonia, about a mile and a half from the 

 town of Argostoli, near the entrance of the harbour, where the 

 shore is composed of freestone, and is low and cavernous, from 

 the action of the waves. 



" The descending streams of salt water are four in number ; 

 they flow with such rapidity that an enterprising Englishman 

 has erected a grist-mill on one of them with great success. I 

 have been informed that it produces him 300/. a year. The 

 flow is constant unless the mouths through which the water 

 enters are obstructed by sea-weed. No noise is produced 

 by the descent of the sea-water, and rarely is any air 

 disengaged ; the streams have been watched during earth- 

 quakes, and have not been found affected by them. It 

 is stated that fresh-water is perpetually flowing through 

 fissures in the rock fi"om the land in the trench which has 

 been dug for the reception of the mill-wheel, and that, when 

 the sea-water is prevented rushing in, then the water in the 

 trench rises higher by several inches than usual, and the water 

 is brackish to the taste. The phenomenon has been long 

 known to the natives. The little information I have obtained 

 respecting these extraordinary currents I owe to my friend Dr. 

 White, surgeon of the Second Battalion of the Rifle Brigade, 

 collected by him when stationed in the Ionian Islands about 

 1840." 



If Mr. Ledger's friend could give us more information it would 

 be most desirable. I am sony I had not an opportunity of 

 examining the mill when I was amongst the islands in i857' 



Gateshead, January 17 R. S. New ALL 



After-images 



Can any reader account for the following interesting pheno- 

 menon : — If I close my eyes in the presence of a strong light, so 

 close that not a ray of light can penetrate the lids — in fact, I 

 may generally place my hands firmly over my eyes — I can see 

 pictures of great splendour, more beautiful than any decoration 

 I have ever beheld, sometimes in the form of some splendid 

 architectural design, most elaborately woiked out ; at others, 

 beautiful landscapes ; again, fine geometrical and other designs, 

 as well as every conceivable form of conventional treatment, 

 such as might be applied to carpets, or other floor decorations, 

 iron-work, &c. I would add that all this is seen without any 

 apparent preconceived action of the will, as sometimes, if I 

 close my eyes with the deliberate intention of seeing any particular 

 object, I am disappointed, though not so frecjuently now as when 

 I first noticed the phenomenon a few years ago. I have sometimes 

 seen designs positively ugly, but as a rule they are most beautiful 

 in form and colour. 



I have visited but few grand and noble buildings, and seen but 

 little of beautiful landscapes, as I am only a humble mechanic, 

 but I take great delight in reading descriptions of such buildings 

 and scenes, and am a true lover of sound, substantial, and 

 elaborate workmanship. J. C. S. 



PROFESSOR TAIT ON THE PARTITION OF 

 ENERGY BETWEEN TWO SYSTEMS OF 

 COLLIDING SPHERES^ 

 CINCE Clerk-Maxwell published, in i860, his f^rst 

 •^ grand investigation on the subject, it seems to have 

 been taken for granted, rather than proved, that in a 



I Abstract of Paper read to the Royal Society of Edinburgti, January i3. 

 " ■ aled by permission of the Council. 



inixture of great numbers of colliding spherical particles 

 of two kinds, the ultimate state would be one in which 

 the average energy of translation is the same for a sphere 

 of either kind. Also that his Corollary, which extends 

 the proposition to a mixture of many systems, is true. 

 Further extensions have been made, the results of which 

 have been considered as irreconcilable with the kinetic 

 theory of gases, at least in its present form. 



So far as I am aware, no really convincing proof of 

 this theorein has yet been given. Maxwell's first proof 

 is so sketchy, and involves so many inadmissible assump- 

 tions, that it cannot be looked on as more than an illus- 

 tration of a truth wJiich his deep insight had enabled him 

 intuitively to perceive. More recent proofs depend so 

 much on a species of analytical verbiage (under cover of 

 which any amount of assumptions may be tacitly intro- 

 duced), that, besides being totally unintelligible to any 

 but specialists, they do not bring full conviction even to 

 specialists themselves. What is required is plain, clear 

 statement, and justification of every step about to be 

 taken, such as will commend it to the careful reader, and 

 leave no doubt on his mind as to what is about to be done, 

 and luhy ; though the mere details of the subsequent 

 necessary calculation maybe beyond him. Nothiirgdoes 

 greater harm to the average reader, in the way of shaking 

 his belief in the results of an investigation, than the use 

 of analysis instead of, or so as to mask, thought. One may 

 make a mistake in evaluating a definite integral, just as 

 one inay make a mistake in adding a column of figures. 

 But when the process of forming the expression to be 

 integrated, or of obtaining the items of the column of 

 figures to be added, is not made fully intelligible, in- 

 credulity is very justly aroused, however we may be 

 inclined to trust the special skill of the mere analyst or of 

 the arithmetician in his proper sphere. 



In seeking such a convincing proof, I have become 

 from time to time suddenly aware of specially dangerous 

 traps which (some almost obvious, others extremely diffi- 

 cult to detect) abound in this particular region of inquiry. 

 .Some of these will appear in what follows. Hence I 

 determined not to be content with anything short of 

 absolutely pointing out the nature of, and the reason for, 

 every step ; so that even those who cannot follow the 

 step itself may fully understand why it has been taken, 

 and be in a position to judge of its legitimateness. 



Limits of space forbid my giving all this in an abstract, 

 so that I must confine myself to a very condensed state- 

 ment. 



For reasons given, we assume the truth of the " error- 

 law " distribution of speeds in any one system of spheres. 

 This will be called the " special " state. 



When two systems are mixed, we assume the mixture 

 to be complete ; and, on account of the small fraction of 

 the whole number of particles (one from each, or one from 

 either, system), which are at any time in collision, and of 

 the perfect fi-eedoin of colli si o?i between any t wo assigned 

 particles (this is a point of special importance), we assume 

 that each systein, by its internal collisions, maintains its 

 own " special " state. Hence in our investigations the 

 collisions of two particles of the same system need not 

 be attended to. Their sole function has been assigned, 

 and we assume that they accomplish it. 



15ut it is most distinctly to be understood that the above 

 assumptions are absolutely necessary to the prosecution 

 of the inquiry in the manner adopted ; and, therefore, to 

 whatever result it may legitimately lead, that result is not 

 to be held as accurate if any of them be departed from. 

 Thus the extensions of Maxwell's Theorem, given by 

 IJoltzmann and others, must not be considered as legiti- 

 mate extensions of that Theorem and its corollaries 

 unless, in the collisions between complex particles, the 

 mechanism of each degree of freedom of any one such 

 particle has perfectly free access for collision with that of 

 the corresponding as well as with that of the non-cor- 



