8o 



NATURE 



[May 25, 1905 



ing mantle or " jacket " of such enormous density and 

 altitude as to contain (as its main constituents) (a) the 

 greater part of the water of the present hydrosphere in 

 the vapour state ; (6) the CO, locked up in the lime- 

 stones and other carbonates of the lithosphcre, as well as 

 that represented by the coal and the living vegetation of 

 the globe ; (c) the hydrocarbons possibly represented by 

 Archsan graphite, together with (d) the halogens (it 

 atomic evolution had reached that stage), including the 

 CI, of the 73 per cent, of the NaCl of the salts of the 

 present ocean. It is conceivable that a vast convection 

 system existed, as the outer zones of the primordial atmo- 

 sphere underwent cooling with consequent condensation, 

 and descended towards the molten globe ; but there could 

 scarcely be contact generally between such cooler portions 

 and the heated molten mass. The conditions would be 

 rather such as are partly illustrated by what a student of 

 physics is familiar with as the " spheroidal state " of a 

 liquid floating on a cushion of steam above a hot plate of 

 metal. Under the enormous pressure prevailing at the 

 surface of the globe in that pre-oceanic stage of its history 

 great quantities of superheated steam and other gases 

 must have been mechanically included, and in some cases, 

 perhaps, occluded, in the hot cruU in the inceptive stages 

 of its development by congelation ; and in such circum- 

 stances, as I suggested seventeen years ago, superheated 

 water in traces would probably enter into the composition 

 of such silicates as hornblende and mica, the two most 

 characteristic of the minerals of the heavier metals of the 

 Archsean gneisses and schists. A year or two later that 

 hypothesis received demonstration from the splendid work of 

 de Kroustchoff (see Nature, vol. xliii. p. 545, also Bulletin 

 de y Academic des Sciences de St. Petersiourg, tome xiii., 

 " Uber kiinstliche Hornblende," by K. von Chrustschoff). 

 So, I fake it, we can understand how such a crust could 

 float on a magma of molten rock material, just as air- 

 charged fragments of pumice or of charcoal float on water, 

 yet sink quickly to the bottom under the exhausted re- 

 ceiver of an air-pump ; or as even a coil of platinum foil 

 (sp. gr. 21-5) can be made to float in water inside a good 

 air pump, as it is pontooned bv innumerable bubbles of 

 distended atmospheric gases previously condensed upon its 

 surface ; or, again, as masses of ' lava slag of large 

 dimensions are seen to float for a time upon the vast lake 

 of liquid rock material in the crater of Kilauea. With 

 tidal action in the magma greater when the moon was 

 nearer the earth than at present, such a thin crust would 

 easily undergo disruption, while portions of it would float 

 ott and be engulfed in the magma. This view, which I 

 propounded some seventeen vears ago. had been antici- 

 pated partly by Zollner, and was "adopted bv the dis- 

 tinguished American geologist. Dr. A. C. Lawson, to ex- 

 plain the phenomena presented bv the enormous inclusions 

 of more basic rock masses in the gneiss of the Rainv Lake 

 region which excited great interest among our leading 

 British geologists at the International Geological Congress 

 in London in 1888. though it seems at the time to have 

 been very imperfectly perceived bv most of them. So far 



^ ."Ju j'"' ""'^ ''f •= ^""^^ '° '"PP°''t D"-- See's contention 

 that the descent of such masses into the magma would be 

 arrested long before they even approached the centre of 

 the sphere: but one feels great diff)cultv in following his 

 argument based on "Laplace's law." for reasons Siven 

 in my former letter (Nature, Mav 4). 



By a slip I wrote, it appears, '' impossibility " for possi- 

 bility in the top line of p. 8 in mv last letter. 



Bishop s Stortford, May i;. A. Irving. 



The Spirit-level as a Seismoscope. 

 A MiscoNCE^PTiON seems to prevail among seismologists 

 1 u uuf °''^^''''°"'' of a spirit-level. A displacement of 

 the bubble IS regarded as conclusive evidence of the tilting 

 ^l '*'.?,. '".st^ment. It should be pointed out, however, 

 that his IS far from being the case. For a second cause, 

 equally effective in producing displacement of the bubble 

 Hnn „?''T"'f'u''"'^!?r''''°" °' "^e instrument in the direc- 

 lon of the tube. The position of the bubble should be 

 taken as indicating, not the normal statical vertical, but 

 NO. T856, VOL. 72] 



the dynamical residual vertical obtained by subtracting the 

 acceleration of the instrument (as a vector) from that of 

 gravity. (I disregard, in this statement, the slight lag 

 due to viscosity.) 



A couple of simple experiments, serving to emphasise 

 this, may be suggested. A spirit-level is suspended in a 

 horizontal position by two equal strings attached one to 

 each end. In one case the strings hang vertically from 

 two hooks ; in the other case they are attached both to 

 one hook. If the level is set swinging in the plane of the 

 strings, then in the first case the bubble will be found to 

 have an oscillatory movejiient relatively to the tube, the 

 tube having linear acceleration but no tilting movement. 

 In the second case the tube has both movements, but their 

 effects exactly neutralise each other, and the bubble re- 

 mains stationary in the tube. The expert waiter (may it 

 be added?) who hurries about with plates of soup has a 

 very effective empirical knowledge of this last case of 

 compensation. 



The motion of the bubble of a level has been brought 

 forward as evidence in favour of the undulatory character 

 of the disturbance producing the motion ; but if the above 

 suggestions are to be accepted, the motion might as reason- 

 ably be urged as evidence of a horizontal disturbance ; the 

 truth being that the instrument is sensitive to both dis- 

 turbances, and is quite ineffective as a means of dis- 

 criminating between them. 



The evidence referred to is contained in the British 

 Association report, igo2 (seismological committee report, 

 p. 72). The view finds acceptance in some recent and 

 authoritative works,' and seems, so far, to have passed 

 unchallenged. G. T. Bennett. 



Emmanuel College, Cambridge. 



A Feather-like Form of Frost. 

 The accompanying photograph shows a form of frost 

 not, I believe, usually seen except at a comparatively high 

 altitude and unsheltered position. This photograph was 



taken on April 22 near the summit of Carnedd Llewelyn, 

 N. Wales (3484 feet above sea-level). These delicate frost 

 " feathers " appear gradually to grow outwards from the 

 rock face on the windward side, and the delicacy of their 

 form is, no doubt, modified in some degree with the vary- 

 ing rate of the wind and the temperature. I have found, 

 in the same district, these " feathers " 9 inches from root 

 to tip; those show-n are about 6 inches long. They form 

 a comparatively solid mass where they touch, but the tips 

 keep distinct, and the whole mass is in reality very brittle, 

 and easily breaks up into small pieces. 



H. M. Warner. 

 44 Highbury Park, N., May 16. 



1 Dutt^n, " Earlhq.iakes in th- Light of the New Seismology." p. 137 ; 

 Davison, "A Study of Recent Earlhquakes," p. 280. >- j/ . 



