1875 .
] On the Expansion of Substances on Solidification .
209 ilurdi ( D. ) Barometrografo Elettro-Magnetico .
8vo .
Napoli 1874 .
Di un fenomeno dipendente dalla di versa densita dell ' Acqua .
8vo .
1874 .
The Author .
Vood-Mason ( J. ) On Indian and Malayan Telphusidae .
Part 1 .
8vo .
Calcutta 1871 .
On a case of Polydactylism in a Horse .
8vo .
1871 .
On Nephropsis Stewarti .
8vo .
1873 .
On Kroyeri .
8vo .
1873 .
On New Species of Phasmidae .
8vo .
1873 .
On a New Genus of Land-crabs from the Nicobar Islands .
8vo .
1873 .
The Author .
Wormald ( Thomas)\#151 ; Obituary Notice .
8vo .
London 1874 .
G. Gulliver , F.R.S. " On the alleged Expansion in Volume of various Substances in passing by Refrigeration from the state of Liquid Fusion to that of Solidification .
" By Robert Mallet , F.R.S. \amp ; c. Received April 28 , 1874* .
The fact that water expands in becoming ice , and that the latter thus floats upon the water , can scarcely have escaped the observation or inference of the acute intellects of a remote antiquity .
Its conditions , when more carefully examined in modem times , pointed out the strange and , as it has been called , anomalous fact that water can be cooled 7 ' or 8 ' below its freezing-point without becoming solid , and that between its maximum density at about 39 ' Fahr. and its freezing-point at 32 ' Fahr. , or within the narrow range of 7 ' Fahr. , it expands in the large ratio of 915:1000 .
Standing thus alone amongst observed phenomena in nature , it seems to have suggested to many experimenters the question whether other bodies when liquefied by heat might not also expand when becoming solid by refrigeration .
I have not attempted to trace with minuteness the history of past inquiry upon this subject , many loose uncertain statements as to which have for at least a century continued to perplex scientific literature .
Reaumur appears to have been the first who gave currency to the statement that cast iron , bismuth , and antimony all expand in consolidating .
The like fact has been alleged or left to be inferred with respect to the following substances by the authorities named:\#151 ; Silver , Persoz .
Copper , Karsten .
Mercury and Gold , as inferred by Nasmyth and Carpenter .
* Read June 11 , 1874 .
See 'Proceedings , ' vol. xxii .
p. 366 .
Mr. R. Mallet on th alleged* Expansion Iron and Furnace-slags , by experiment of Heunter and Snelus , as quoted by Nasmyth and Carpenter .
But of this list the only body , in addition to water , that really appears proved to expand in consolidating is bismuth ; and even this the author cannot affirm upon the basis of his own experiments , but accepts the fact , at least provisionally , as true upon the uncontradicted statements of many chemical authors , and upon the positive assurance which he is permitted to mention by Dr. John Tyndall that he is satisfied of its truth .
" With respect to all the others , it is the object of this communication to show that the evidence in support of the alleged fact of expansion by refrigeration is illusory and insufficient , and to offer with respect to cast iron , and also with respect to iron furnace-slags , experimental proof of the untruth of the statement .
Certain connected but only collateral facts , having regard to so-called anomalous changes of volume due to temperature , will not be referred to here\#151 ; such , for example , as the anomalous expansion of Rose 's fusible metal , which expands progressively , like other bodies , till it attains the temperature of 111 ' ; it then contracts rapidly by added heat to 150 ' , when it is densest ( Grraham 's ' Elements , ' vol. i. , and Qmelin 's 'Handbook ' ) , the circumstances being here probably due to the successive segregation in the mass of alloys differing from each other in constitution , dilatability , and fusing-points .
Or , again , the facts observed with respect to the expansion or contraction in volume shown by certain salts when crystallizing from their solutions , the whole of the conditions as to which have not been as yet made quite clear .
The statement that antimony expands in consolidating , as made by Reaumur , has been negatived by Marx .
The like statement with respect to silver and copper appears to rest on no better foundation than the observation as stated by Persoz , " that pieces of solid silver float upon the melted metal , showing that silver expands in solidifying like water " ( G-melin 's * Handbook , ' vol. vi .
) .
As to g'$\#174 ; i there appears no authority whatever for its expansion on consolidation .
Mr. Nasmyth has included it in his catalogue merely on the vague in-ference that , like silver and copper , it " exhibits surface-converging currents in the melting-pot like those depicted by him for molten iron " ( ' The Moon , ' Nasmyth and Carpenter , p. 24 ) , which , as we shall see further on , affords no grounds for conclusion on the matter .
Reaumur 's statement with respect to cast iron appears to have rested upon nothing more than the fact that he had observed certain pieces of cold cast iron to fldP upon cast iron while in fusion .
Until lately this subject generally attracted but little attention , for it had very few , and these mere technical , applications ; and to the higher physicist they presented but little interest , because the loosely stated facts , even if accredited , did not in the slightCP degree tend to elucidate or explain the remarkable and perhaps stiff is\#174 ; lated facts as to water and ice .
Accordingly , with little or no exam1 " nation , the statements given for facts by the older authorities have b\#174 ; B j of various Substances on Solidification .
:cepted and become current from book to book of authors up to the \#166 ; esent day , as when Dr. T. Thomson ( ' System of Chemistry , ' vol. i. 375 , 5th edit .
) lsays of cast iron that " it contracts considerably when comes into fusion , " or that of Kerl ( ' Metallurgy of Iron , ' Crookes and ohrig 's translation , vol. ii .
p. 291 ) , that cast " iron occupies a smaller \gt ; ace after cooling than when in the liquid state ; it contracts in such a anner that , at the commencement of its solidification , it first expands so i to be able to fill up the smallest depressions and cavities of a mould , but fter solidifying it contracts"\#151 ; a loosely worded statement , which in irious forms may be found in a great number of authors upon metal-irgy and technology .
So likewise the statement often repeated , that ie value of antimony in type-metal consists in its causing the latter to cpand upon consolidation and so perfectly fill the matrix , is presented , ) far as the author 's reading goes , without the slightest experimental roof of its truth , and appears to rest simply upon Reaumur 's statement ith respect to antimony itself , which , as already mentioned , has been introverted by Marx .
This subject , however , has now assumed greater nportance , since it has recently been made by Messrs. Nasmyth and Carpenter the foundation upon which they rest their theory of lunar olcanic action , as presented to us by the surface of our satellite ; and the bject of the present communication is to show that , as regards the two lost pertinent of the substances adduced by these authors , viz. cast iron nd iron furnace-slag , the facts entirely fail in support of their theory .
First , then , as to cast iron .
It is not a fact that all cast iron in the olid state will float upon all cast iron in liquid fusion , though such might \gt ; e inferred from the broad and loose statements of authors .
Even in the united form in which the statement is made by Nasmyth and Carpenter\#151 ; iz .
" that when a mass of solid cast iron is dropped into a pot of molten ron of identical quality the solid is found to float persistently upon he molten metal , so persistently that when it is intentionally thrust to he bottom of the pot it rises again the moment the submerging agency 3 withdrawn " ( ' The Moon , ' p. 21)\#151 ; is not quite exact .
It is a fact that certain pieces of cast iron in the solid and cold state will float on certain descriptions of cast iron in liquid fusion ; but whether he solid pieces shall float or not float in any given case is dependent at east upon the following conditions , and probably upon others not yet iscertained :.\#151 ; 1st .
Upon the relative specific gravities of the solid and of the fused ; ast iron both referred to the temperature of the atmosphere .
Under the commercial name of cast iron is comprehended a wide range of compounds of iron with other substances , which compounds differ greatly in their physical as well as their chemical qualities , and have a range of specific gravity of from nearly 7*7700 for the whitest , most rigid , and dense , down to little more than 6*300 for those which are darkest , softest , and most porous .
The total dilatation at the fusing-point of the denser Mr. R. Mallet on the alleged Expansion .
cast irons is known to be somewhat greater than that o\#163 ; the less dense ; but as the increase in volume may not be sufficient to equalise the specific gravity of a very dense iron when in f usion with that of a very light iron when cold , so it is obvious that a piece of cold cast iron might be so selected in reference to its specific gravity , as referred to that of another sort of cast iron in fusion , that the former should either sink or swim upon the latter by mere buoyancy , were that free to act alone .
2nd .
Assuming the cold and the molten cast iron originally identical in qualities , whether a piece of the former shall float or not float upon the latter depends not only upon buoyancy as above , but also upon the form of the piece of cold metal\#151 ; that is to say , on the relation , all other things being the same , that subsists between its volume and its surface .
3rd .
The force , whatever be its nature , which keeps the piece of cold cast iron floating is of sufficient energy to overcome a considerable want of buoyancy in the cold iron under certain conditions , so that it may float upon molten cast iron whose specific gravity , as such , is much less than that of the colder iron which floats upon it .
Messrs. Nasmyth and Carpenter assume , without any sufficient proof , that solid cast iron floats on liquid iron of the same quality in virtue of buoyancy alone , and proceed to state that " the inevitable inference from this is that in the case of cast iron the solid is specifically lighter than the molten , and therefore that , in passing from the molten to the solid condition , this substance undergoes expansion in bulk " ( ' The Moon/ pp. 20 , 21 ) .
I proceed to prove that this view is altogether contrary to fact .
The determination of the specific gravity of cast iron in its molten condition is a problem of considerable difficulty , and can only be solved by indirect means ; we cannot ascertain its specific gravity by any of the methods ordinarily employed , nor can any are- _ ometric method be used , as any hydro-meter or solid of known specific gravity at common temperatures , when dipped into liquid cast iron , changes its volume as well as gets encrusted with adherent cast iron or its oxides , \amp ; c. By an indirect method , and by operating upon a sufficiently large scale to eliminate certain sources of error , the specific gravity of molten cast iron may , however , be approxi--mately ascertained with considerable accuracy .
The method adopted by the author was as follows : \#151 ; A conical vessel , of the form shown in figure 1 , was formed of wrought-iron plate by welding up only , of various Substances on .
13 walls of the vessel being about S in .
in thickness.1\#174 ; It was per-jstly smooth in the inside , and the plane of the lip of the open ick was carefully made parallel to the plane of the base .
This i3sel weighed , when empty , 184*75 lbs. avoirdupois .
The orifice of 13 neck being levelled as the vessel stood upon the platform of the i :ighing-apparatus , it was filled up to the exact level of the neck with iter at a temperature of 60'*5 Fahr. , and again weighed .
Deducting the .light of the empty vessel , the weight of its contents of water was found 1 be 94*15 lbs. avoirdupois .
From the known volume and weight of the iperial gallon of distilled water , the capacity of the vessel was therefore : 60 ' Fahr.=2605*5 cubic inches .
As a check upon the results , both ; to weight and capacity , the water was measured into the vessel \gt ; m accurate glass standards of volume .
The water employed was that )m the well at Messrs. Maudslay , Sons , and Field 's Engine Works , tmbeth , where these experiments were conducted , and to whose liberality e author owes the means of having performed them .
The specific avity of this well-water did not very materially exceed that of distilled iter , being about 1*0004 ; but if we apply the necessary correction , the light of the contents of the iron vessel of distilled water at 60 ' Fahr. 94*112 lbs. avoirdupois .
The vessel being emptied , carefully dried and irmed , and stood upon a hard rammed bed of dry sand with its neck rfectly level as before , was now filled perfectly level to the brim with olten cast iron .
As the temperature of the vessel itself rapidly rose by ntact with the large mass of molten iron within it , and by its dilata-\gt ; n had its capacity enlarged , so the top surface of the liquid cast in within it rapidly sank , fresh additions of molten iron being constantly id to maintain its top surface level with the brim .
This was continued itil the whole of the exterior of the vessel was seen to have arrived at a lar yellow heat , beyond which no increase to its temperature took place .
$ about twenty minutes after the molten iron was first poured into the ssel , this point was reached , the feeding in of additional iron being dis-ntinued a few minutes previously .
The whole being left to cool for ree days , the vessel full of the now cold and solid cast iron was again lighed ; on deducting , as before , the weight of the empty vessel , the weight the cast iron which filled it was found to be 645*75 lbs. , which , with rtain corrections to be yet noticed , was the weight of cast iron which , len in the molten state , was equal to the capacity of the conical iron ssel in its expanded state due to its exalted temperature .
We have \gt ; w to determine what was the capacity of the vessel in this expanded a-te .
The temperature at which cast iron melts may be admitted as out 2400 ' Fahr. ; but as iron tapped from the cupola is always above \lt ; melting-point , we may admit that it was poured into the vessel at \gt ; 00 ' or 2700 ' Fahr. , the surplus heat in the cast iron , whose mass was out four times that of the wrought-iron vessel which contained it , being ven off in the first instance to heat the latter .
The temperature at which Mr. R. Mallet on the alleged Expansion wrought iron presents to the eye a clear yellow visible in daylight may in accordance with the views of most physicists , be taken as between the fusing-points of silver and of gold , or at 2000 ' Fahr. The mean coefficient of linear dilatation for 1 ' Fahr. of wrought iron has been determined between the limits of zero and 212 ' by Laplace , Smeaton , Troughton , and Dulong , the average of the four being 0*00000699 for 1 ' Fahr. ; and this is certainly below the truth for the whole range of temperature up to fusion , as the rate of expansion of all fusible bodies appears to increase with the temperature .
Rinmann has determined the linear dilatation of a bar of wrought iron , when raised from 60 ' Fahr. to a white or welding heat , to be of its length , or *0125 ; and taking the total range of temperature here at 2400 ' , we have a mean coefficient of linear dilatation = 0*0000052 for 1 ' Fahr. This is a still smaller coefficient than the preceding ; the author has , however , preferred to adopt it in order to avoid any pretence to exaggerate in his own favour the results arrived at .
Applying , then , Rinmann 's coefficient to the dimensions of the cone at 60 ' Fahr. , and to its temperature ( 2000 ' Fahr. ) when at the maximum , we are enabled to deduce the true capacity of the cone when expanded to the utmost and filled with molten iron , viz. =2691*77 cubic inches .
The iron conical vessel was now cut off by a circular cut at the base and another up and down the side of the cone , and separated from the conical mass of iron that had filled it ; the interior surface of the iron vessel was found in several places about the lower part of the cone in perfect contact with that of the cast iron which had filled it ; but in other portions very slightly distant from it , as judged by the sound of ' a hammer upon the sides of the vessel before it was cut off .
The cast iron was not adherent to the vessel anywhere .
The cast-iron cone being thus laid bare , had a Y-shaped piece cut out of it ( in the " slotting''-machine ) , as shown in fig. 2 .
by two planes , each passing through the axis and meeting at an angle of about 60 ' .
The conical mass proved perfectly sound and free from cavities or blow-holes anywhere , except very near the summit or neck , where there was found to be a hollow or cavity accidentally left during the feeding ( as above described ) .
By measurement the volume of this cavity was found to be = 5*5 cubic inches ; assuming this cavity filled with iron of the same quality as the cone , the weight of the latter would be increased by 1*43 lb. , making thus the corrected total weight of the solid cour of cast iron = 647*18 lbs. From the wedge-shaped piece cut out from the cone at half its altitude , and about halfway between the axis and cir cumference of the sector , a piece was cut out , the specific gravity of which Fig. 2 .
of various Substances on Solidification .
teen by the usual methods , proved to be 7*170 , which may be taken as td mean specific gravity at 57 ' Fahr. of the whole of the cast iron that t ed the cone .
Reverting now to the conical vessel which contained at l ' Fahr. 94*112 lbs. of distilled water , its capacity being 2605*5 cubic i ; ;hes : this capacity was enlarged by expansion when filled with molten .\gt ; n to 2691*777 cubic inches , so that the conical vessel when cold , if it id had the same capacity as when filled with liquid iron , would have ntained 97*224 lbs. of distilled water .
We have now all the elements jeessary for calculating the specific gravity of the cast iron which filled e cone in its molten state , because we have the actual weights of equal dumes of distilled water and of molten iron .
The final results , then , .e , that whereas the cast iron which filled the cone had when cold ( 57 ' ahr .
) a specific gravity , as above given , =7*170 , the same cast iron in 3 molten state , as poured into the cone , had a specific gravity of only 650\#151 ; in this case thus proving that the density of cast iron in its liquid .ate is not greater but , on the contrary , very much less than that of the ime cast iron at the temperature of the atmosphere .
The quality of ist iron employed in this experiment was the fine , bright , close-grained letal usually employed by Messrs. Maudslay , Sons , and Field for their ngine-castings , and consisted of \ Best scrap*\#151 ; all by weight .
t may be taken as a typical or medium example of all good grey cast irons , have not been enabled to repeat this experiment with the white , rigid , nd crystalline cast irons , such as are employed for projectiles and other nrposes ; but as it is a recognized fact amongst iron-founders that these rons expand in the range of temperature between solidity and liquidity 'luch more than do grey irons , so we may justifiably conclude that the decrease of specific gravity by fusion of these hard cast irons would be in even i greater ratio than that shown by the above experiment on grey iron ; and generally the author feels himself justified in concluding that it is not rue that any cast iron is denser in the fused than in the solid state .
Cold cast iron , therefore , does not float upon liquid cast iron of the same quality by reason of its buoyancy , but in virtue of some force which tends \#166 ; o keep it upon the surface of the molten metal in opposition to a very jonsiderable want of buoyancy or tendency to sink by greater density on she part of the solid iron , which is , by the preceding results , of its veight , whatever that may be , and is probably even greater than this in he case of hard white cast irons .
The author 's chief object has been thus far rather to prove that the cause assigned by the writers already mentioned is not the true cause of the floating of solid upon liquid cast iron of the same quality .
What is the nature of the force which pro* Disused and broken-up castings .
VOL. XXIII .
s Mr. R. Mallet on the alleged Expansion duces this curious phenomenon and often in direct opposition to gravity is a different and a much more delicate and difficult inquiry , which he must leave to physicists to fully investigate .
The following experiments , however , may be placed on record as tending to afford some little dawn of light upon the subject .
The following experiments were made with pieces of iron cast from cast iron of the same quality as that which filled the experimental cone , placed upon or immersed in molten cast iron of like quality with themselves , so far as such can be secured by " tapping " at nearly the same time from the same cupola charged with the same materials .
Before proceeding to describe these , it will be necessary to deduce from the cone experiment a mean coefficient of total cubic dilatation for the whole range between 60 ' and 2400 ' Fahr. for the grey cast iron employed in these experiments .
The total dilatation was , as we have seen , such as reduced the specific gravity of the cast iron when cold ( = 7*17 ) to 6*65 when in fusion .
The cubic dilatation was therefore in the inverse ratio of these numbers , or as 1000 :1078 ; and dividing this increase in volume by 2340 ' Fahr. , the total range of temperature , we obtain foi the mean coefficient of cubic dilatation of this grey cast iron for 1 ' Fahr. = 0*0000333 , or approximately for its mean coefficient of linear dilatation 0 -''30333 = 0*0000111 .
These coefficients are nearly double those obtained by Roy and by Lavoisier for a range of temperature of 180 ' Fahr. , viz. between 32 ' and 212 ' , which is quite what we should expect , as the coefficient of dilatation in all bodies increases tfsvith the temperature .
" We have seen from what precedes that twci forces at least are concerned in the phenomenon of cold cast iron floating upon the same when liquid , viz. A. Buoyancy or its opposite , dependent upon the relation between the actual specific gravity of the cold metal and that of the liquid metal upon which it is placed , and whose absolute power for any given difference of specific gravity depends upon the volume only of the floating mass .
B. A repulsive force of some kind tending to repel the surfaces in contact of the hot and cold metals .
Whatever be the form of the floating solid , this repulsive force can only be effective in producing flotation upon such surfaces of the floating solid as are parallel to the surface of the liquid metal , or at least so circumstanced that repulsions upon one surface , or part of a surface , are not equilibrated and nullified by repulsions upon others in the opposite direction .
Thus if a parallelopiped float with one of its surfaces parallel to that of the liquid metal , the repulsions upon its immersed vertical sides , taken two and two respectively , are in opposite directions , and therefore nullified , and the bottom or horizontal surface is alone effective in producing flotation .
So also if a cylinder float with its axis horizontal , the ends are ineffective , as is also all that portion of the cylindric surface immersed which is above the level of the horizontal diameter of the cylinder .
of various Substances on Solidification .
217 'hese preliminary explanations will enable us better to interpret the !
owing experiments .
Experiment 1 , An irregular piece , believed to be of hard and dense a# iron , and also a ball of about 2^ in .
diameter , believed to be of close-med grey iron : both sunk to the bottom when thrown into the ladle ofiquid iron , and remained for some time at the bottom ; both , however , uopeared upon the surface when they had acquired a temperature suffi-ut to have fused off portions of their respective masses .
In every fresh-lined ladle of liquid cast iron there are circumferential a ending and central descending currents in the metal , produced by the gaes evolved from the lining , as hereafter fully explained .
It is no doubt lefly to these ascending currents that the heated ball in Experiment 1 oed its ascent to the surface ; for if the heating took place in perfectly nionless cast iron , there seems no reason why the place of the sunken hi should change up to the moment of complete fusion* .
] iJaeperiment 2 .
Two parallelopipeds , each 2 " x 2 " x 6 " , were cast of close \lt ; y j iron ; one of these was placed cold upon the surface of a large ladle 0 liquid iron of like quality ; the other was heated as hot as it would n , r without distortion , viz. to nearly a bright yellow heat , in a forge-fire , a l then placed upon the surface of the liquid metal .
Both pieces floated , al , as nearly as could be judged , both to the same height above the laid , namely 0*1808 in .
The volume of the cold piece being 24 cubic 1 hes , the ratio of the immersed to the emergent portions was as 9*6 to the effective surface upon which the repulsive force could act in pro-tfing floatation being 12 sq .
in .
Assuming that the heated piece has tm raised from 60 ' Eahr .
( the temperature of the cold piece ) to 2000 ' hr .
, and applying the mean coefficient of cubic dilatation as above given t this range of temperature , viz. 2000'\#151 ; 60 ' = 1940 ' Eahr .
, we find that \gt ; volume was enlarged to 24*75 cubic inches , or of the volume ten cold ; and taking the specific gravity of the cold piece to have been 17 ( see ante ' ) , that of the hot piece would be reduced to 7*10 ; the effec-e repellent surface was slightly enlarged in the hot piece , and the im-irsed volume was to the emergent volume as 9*66:1 .
The buoyancy .
the heated piece had been increased , or , more correctly , its negative nyancy had been decreased , as compared with that of the cold piece , but , t it has sunk deeper into the liquid iron in proportion to their re-ective volumes .
We may therefore be justified in concluding that the pellent force which kept both pieces afloat is diminished in energy in me proportion as the difference in temperature between the liquid metal id the piece floating upon it is diminished , and that where the liquid id the floating pieces are alike in quality of metal , both the negative loyancy and the repellent force must both disappear at the instant that ie floating piece itself becomes liquid by heat abstracted from the molten 3tal .
* All passages printed in brackets take date from 20th December , 1874 .
s 2 218 Mr. R. Mallet on the alleged Expansion Experiment 3 .
Two cylindric pieces of the same grey cast iron and c the same diameter ( =2-375 " ) were gently placed with their axes hot zontal upon the surface of the molten iron , the one being at 60 ' Fabr the other at about 300 ' Fahr. ; they both floated with a segment of tb cylinder whose versed sine was 0-31 in .
emergent .
The volume of eithe cylinder was 22-15 cubic inches , and the emergent was to the immerse volume as 1 : 8-4 .
The effective repellent surface in each case ( or cylii dric surface below the horizontal diameter ) was 18-65 sq.in .
; but if u suppose , as in fact we have done , that the repellent force , whatever be ii nature , acts everywhere perpendicularly to surfaces of contact of tl solid and liquid , then the effective repellent surface here is only the diffei ence between the immersed surfaces of the cylinder below and above tl horizontal diameter , or 9*3 sq .
in .
From this we may perhaps concluc that the repellent force is mainly dependent upon the extreme upper par of the range of temperature between the liquid and the cold body , ina much as an augmentation in temperature of the latter of 300 ' , or aboi of the entire range between solidity and fusion of the cast iron , prt duces no very sensible alteration in the tendency to float of the pieces .
Experiment 4 .
Three circular disks of the same grey cast iron , each 6 " diam. by 0-375 " in thickness , were provided each with a slender ire wire eye , cast into the centre of one surface , so that by a hooked wi they could be gently laid upon the surface of the liquid iron of their ov quality .
The lower surface and edge of one disk were left as it came cles from the sand , those of another were rusted by wetting with solution sal-ammoniac , and those of the third were ground smooth and polished1 the grindstone .
When the three disks were in succession laid upon t !
surface of the molten iron , they all floated alike as nearly as could judged , each sinking to one half the thickness of the disk , so that fe immersed was to the emergent volume in the ratio of equality .
|P may conclude from this that the condition of the metallic surface of t solid cast iron has no material influence upon its flotation .
Experiment 5 .
Two circular disks , provided with eyes as in experiment were prepared , the one being 6 in .
in diam. by 0-375 in thickness , and t other 3 in .
in diam. by 1*5 in .
in thickness .
The respective volumes these two disks are the same , but the circular flat surfaces respectiv* are as 4 to 1 .
The surfaces of the two disks being as they came from t sand-mould , they were placed gently upon the surface of the molten iro both floated with the same portion in altitude emergent .
The larj and thin disk had , as stated in experiment 4 , its emergent and immers volumes in the ratio of equality [ or the emergent was to the whole volu .
as 1 : 2 ] .
In the smaller and thicker disk , the emergent volume was the immersed volume as 1 to 7 .
[ Or the emergent volume was to to volume as to 1 : 8 ; but 2 : 8 : : 1 : 4 , or the emergent volumes are to t total volumes in each case respectively proportionate to the lower or ; pellent surfaces of the disk.j of various Substances on Solidification .
ow the effective repellent surfaces are here those of the lower circles p\gt ; fie respective disks , and these surfaces are to each other in the ratio -.I ( the larger ) to | .
Whatever be the nature , therefore , of the repel-i\gt ; .
} force , it seems to be proportionate to some function of the effective *uace as already defined , and not to the immersed volume of the solid i'fi iron which floats upon a liquid less dense than itself .
1 all these experiments the mass of the molten cast iron was large in n portion to the pieces placed upon it , and the surface was kept by eare-f uskimming almost perfectly free from scoriae or oxide .
A good deal .'i ifficulty exists in observing the phenomena in such experiments as rbe , owing to the glare and heat of the molten metal .
Whatever light ' the five experiments may throw upon the nature of the force which nduces flotation , the subject must as yet be viewed as very incomplete . .
re are some facts of which no complete explanation can be offered 'i iout further experimental study ; such as , for example , that a piece of .
r. cast iron which floats on liquid iron of its own quality if forcibly l ist to the bottom and rapidly and at once released , rises again rapidly .
the surface with all the appearance of a buoyant body , which it eerily cannot be .
'rom what precedes , however , we may summarize as follows:\#151 ; f F be the force which keep the solid iron floating , B the buoyancy -iof the solid piece , and R the repellent force , then , in the case of a me floating upon molten iron of its own quality , B is always negative , ai F=R\#151 ; B , the value of R for any given case depending upon the qctive surface of the solid , and that of B upon its volume , both being edified by the initial difference in temperature between the solid and i\#187 ; lid metals .
In the case of the solid being placed on liquid cast iron 1 ering in quality from it , B may be either positive or negative , and R sd dependent upon the conditions already stated .
Hence in any such -e we may have F=R\#151 ; B or = R+B .
Chese conditions kept in view may clear up many phenomena at first atarently anomalous .
However feeble may be the ascending currents , above referred to , t m the floating disks in experiment 5 , their effect must be viewed as nportionate to the lower surfaces , and therefore proportionate to ts repellent force , and as possibly adding , though slightly , to its eect .
] the following experiments were made at the Royal Arsenal , Woolwich , Vjh a view to ascertain whether any sensible expansive force could be Cognized as due to the enlargement in volume by consolidation of a Clerical mass of cast iron:\#151 ; Two spherical bomb-shells , each of about " in diameter and 1"*5 in thickness , whose external orthogonal dia220 Mr. R. Mallet on the alleged Expansion meters had been carefully taken when at atmospheric temperature ( about 53 ' Eahr .
) , were both heated in an oven-furnace .
One of these having been thus heated , but not to a very bright red , was permitted gradually to cool again , and its final dimension when cold noted .
The other shell was withdrawn from the oven when at a bright red heat and immediately filled to a little above the inner orifice of the fuse-hole ( viz. to the line a b , fig. 3 ) with molten cast iron , the quality of this being the very dense mottled grey iron smelted at Elswick Works from the Riddesdale ores , and used in the arsenal for casting projectiles .
The fuse-hole was closed by a screw-plug , which , however , did not reach within an inch of the surface of the molten metal , and the whole surrounded by a sheet-iron screen to keep off currents of air , was allowed to cool -Fig .
3 .
gradually , the dimensions being taken of the sphere as it cooled and contracted at intervals of half an hour until it had become cold .
The enveloping shell was then cut through by the lathe in a great circle at right angles to the axis passing through the fuse-hole .
One of the halves of the shell being detached , the interior surfaces of both hemispheres were found in perfect contact with that of the ball of iron they had contained , but no elastic tension seemed to exist in the shell .
The ball of iron was drilled into and split by steel taper plugs , and sections of it exposed passing through the diameter in a line with the axis of the fuse-hole .
There was no large cavity or " draw " anywhere in the interior , but there were two very small irregular cavities very near the fuse-hole ; and the central portion of the mass embraced by an imaginary sphere of about 3 " in diameter , as shown in fig- 3 .
proved to be " spongy " and granular , as compared with the very dense and close-grained iron that constituted the remainder of the ball .
The following Table shows the course of contraction in dimensions of the filled shell and also of the empty shell in their progress of cooling ) and the change in dimensions and in volume of the filled shell are shown by the curves ( figures 4 , 5 , 6 , p. 222 ) of various Substances on Solidification .
Time .
Diameter , filled shell .
Diameter , empty shell .
11.30 12.30 Cold. .
Put in oven-furnace ( shell to be filled ) ii ii ( empty shell ) 9-850 9-843 L2.15 W ithdrawn from furnace 10-020 L2.55 11 11 11 .
* . . .
After filling with iron , diameter was 4 l 10-030 10-040 9-960 L2.50 11 ii 11 11 ii 10-040 1.20 11 ii 11 11 ii 10-020 9-955 1.50 11 ii 11 11 ?
i 10-000 9-950 2.15 11 ii 11 11 ii 9-995 9-875 2.45 11 ii 11 11 19 9-980 9-865 3.15 11 ii 11 11 11 9-978 9-860 3.45 11 ii 11 11 11 9-976 9-855 4.15 11 ii 11 11 11 9-975 9-854 4.45 11 ii 11 11 11 9-973 9-862 5.15 11 ii 11 11 11 9-970 9-852 5.45 11 ii 11 11 11 9*968 9-851 6.15 11 ii 11 11 11 9-965 9-851 6.45 11 19 ) ) 11 11 9-964 9-851 7.15 11 11 11 11 , 11 9-964 9-851 7.45 11 11 11 11 9-963 9-851 8.15 11 11 11 11 11 9-962 9-851 When cold 9-960 9-851 The object of heating and cooling the empty shell was to ascertain lat amount , if any , of permanent enlargement it might suffer , it being i veil-known fact that all solids of revolution of cast iron , and generally \#187 ; all metals of sufficient rigidity , become permanently enlarged by being ated red-hot and permitted to cool .
This arises from the fact that e outer couches of the solid ( a sphere for example ) are the first heated d expanded , and have to draw off more or less from the less-heated iss within .
Tangential thrusts and radial tensions are thus produced the material of the outer couches which disappear , or even become versed , as the progress of heating reaches the interior of the mass ; but the subsequent cooling the entire train of forces is reversed , the ex-rior couches lose heat by dissipation first , and have to accommodate by ngential tensions their dimensions to the still hotter interior , the final suit being that when the whole has cooled the dimensions are greater than fore the solid was heated .
A 32-lb .
spherical shot , which is rather more an 6 inches in diameter , can be thus permanently increased of \amp ; n ch in diameter by a single heating .
It is obvious that the increase will \gt ; much less in a spherical shell than in a solid sphere , and the less as of various Substances on Solidification .
e shell is thinner* .
On inspecting the Table and curve fig. 5 , it will !
seen that the empty shell had its diameter thus.permanently enlarged .
0*008 of an inch ; and had it been heated to as high a temperature as the led shell , we may allowably conclude that this enlargement would have ached 0*01 of an inch .
The filled shell has had its diameter increased r the decimal 0"*11 ; and if we deduct from this the amount of perma-snt enlargement due to heating only , equal to that of the empty shell , e have the decimal 0*11 \#151 ; 0*01=0*10 , which has to be otherwise ac-junted for .
This shell was at a bright red heat visible in clear daylight hen filled with the liquid iron , which occupied the spherical cavity and tout 0*43 in height of that of the fuse-hole .
The temperature of the hell visibly rose by the heat communicated from the liquid metal , and in 0 minutes after it was filled had attained its maximum , the surface being aen at a bright yellow heat in daylight when the first measurement of nlarged diameter was made .
The successive measurements were taken are orthogonal diameters in the direction normal to the fuse-hole by leans of finely graduated steel beam calipers capable of being read to *002 of an inch or even less ; the dimensions set down in the Table are he means of each pair of orthogonal diameters .
The shell was thus teated at the commencement , and before consolidation of its liquid con-ents had taken place to any considerable extent , to within probably 200 ' \#187 ; r 300 ' Fahr. of the temperature of the cast iron within .
The shell and fcs contents are therefore at the commencement very nearly in the same .ondition as though the whole were a sphere of molten iron without any nor or less rigid envelope , if such could exist .
Reverting to what has \gt ; even said above as to the train of forces called into play in a cooling sphere , let us consider what has taken place here .
As the heat is dissi-aated from the exterior of the molten mass , being transmitted through he shell , on ejcoucheafter another of the molten metal in contact with the inner wall of the shell consolidates , the thickness constantly advancing towards the interior , where the metal is still liquid .
If each of these touches in consolidating expanded in volume , such expansion must conspire , with the contraction constantly going on by the abasement of temperature , to produce compression in the central and as yet unsolidified portion of the mass .
If , on the contrary , each as it solidifies contracts in volume ( and , as is the fact , by a larger coefficient of contraction for equal small ranges of temperature before and after solidification ) , then the effect musiL.be that , after the solidified crust has attained a certain thickness and sufficient rigidity , the further progress of contraction of the central portions as they successively solidify must be met by their tending to draw off from the solidified shell , or , in other words , by a drawing-off from each other of the particles of that central portion of * For a more ?
complete-analysis""of the complex strains brought into play by expansion and contraction in the heating and cooling of metallic solids of revolution , the author may refer to his paper " On the Coefficients T\#171 ; and T , .
in large masses of Forged Iron , published in the Minutes of Proceedings Inst. C. E. London , vol. xviii .
p. 299 .
Mr. R. Mallet on the alleged Expansion the sphere which last solidifies .
Now the latter is exactly what has happened : a portion of the exterior and first solidified crust , reaching about an inch and half inwards from the interior of the shell , was found to have a specific gravity of 7-150 at 57 ' Fahr. , while a portion taken close to the centre of the sphere had a specific gravity of only 7-037 ; and this specific gravity would have been still lower ( Or , in other words , the central part of the sphere would have been still more " spongy " ) had it not been fed by drawing downwards a portion of the liquid iron which partially filled the fuse-hole , the portion so drawn down being estimated by the volume of the cavities left at 0*400 of a cubic inch ; so that but for this the specific gravity of the central spongy sphere taken at 3 " diameter would have been reduced to 6-776 .
If we reduce this central spongy mass of 3 " diameter and of the last-mentioned specific gravity to a density as great as that found for the exterior crust , namely 7*150 , the sphere of 3 " diameter would be reduced to one of 2"-138 ; and it is easy to see that in that case the external diameter of the whole sphere of metal and of the containing shell would have been less in a corresponding proportion , and that thus the final dimensions of the shell would have returned to what they were at the commencement , less the permanent enlargement , as measured by that of the empty shell .
If there existed , on the other hand , any sensible expansion in volume of the metal in consolidating , not only would a central " spongy " portion be impossible and the central be the densest part of the whole sphere , but an enlargement of the entire mass and of the covering shell stretched by it must have occurred , so large as to be wholly unmistakable .
[ The importance of the facts elicited from this experiment cannot be too forcibly laid before the reader .
Had the sphere of molten iron , losing heat from its exterior , expanded in volume as couche after coucht it solidified from the exterior , the solidification constantly advancing in-wards , then the central portions of the sphere when ultimately solidified must be found to be the densest portions of the whole mass ; the opposite of which was found to be the fact , the central portions of the experimental sphere being , as stated , the least dense portions of the whole mass .
This alone seems conclusively to negative the supposition of any expansion in volume in cast iron in consolidating .
On examining the curve fig. 4 in connexion with the Table , it will be remarked that between the hours 1.50 and 2.45 there is an irregularity in the progress of contraction which shows itself by a hump upon the curve which might be assumed to indicate a less rate of contraction within this epoch ; and \#187 ; might be further assumed that this apparent reduction arose from the conjoint action of general contraction and partial expansion operating together within some part of the mass ; but this view , which the write ?
believes would be entirely incorrect , appears sufficiently negatived by th\#174 ; following considerations:\#151 ; of various Substances on Solidification .
I 1 .
Between the hours 1.50 and 2.45 but one caliper measurement ns made , namely at 2.15 , and upon this one measurement both the I tistence and the amount of this anomalous part of the curve depend .
v,.i error in this single caliper measurement amounting to 0*006 of an :ch was sufficient to have produced it ; and as the limit of reading of x e beam calipers was to a limit of 0*002 or possibly 0*001 of an inch , mistake in the measurement at 2.15 , or a misreading of only the decimal 04 or *005 at most , is sufficient to account for the anomaly .
2 .
The hump on the curve does not necessarily indicate expansion , and om the early time of its occurrence , viz. only 1 hour 25 minutes from ie commencement of cooling , it seems highly improbable that it could \#163 ; rise from partial expansion then commencing , while as yet a very large roportion of the entire mass must have been still liquid .
, 3 .
If this anomalous part of the curve were really due to expansion , it lust have much more extensively affected the lower prolongation of the jrve , and have shown itself there in a form that would have unmistakably eclared its origin .
4 .
On examining the curve fig. 5 a slight anomaly may be remarked l l the rate of contraction of the empty shell , due no doubt to some light error in the third measurement , or that at 1.50 p.m. In this in-tance it would be impossible to ascribe the anomaly to expansion of any 1 ort . !
The dotted line A ( fig. 4 ) may therefore be viewed as completing the ' urve of contraction .
The curve fig. 6 , representing the volume of the filled shell at successive spochs of cooling , is deduced from the Table ( p. 221 ) , assuming the successive volumes to be proportionate to the cubes of the diametric measure-nents , the curve being a mean drawn through the several points of observation .
] The supposition upon which Messrs. Nasmyth and Carpenter 's theory rests may be divided into two distinct propositions .
1st .
That cast iron is of greater density in the molten than in the solid state .
2nd .
That cast iron in the act of consolidation expands in volume .
These propositions are not identical , although the second is involved in the first .
The first proposition has been already disposed of , and the last recorded experiments appear conclusively to disprove the second .
The phenomena described by Messrs. Nasmyth and Carpenter , and their explanation of the circulating currents observable in large and nearly cylindrical ladles of molten iron , appear at first sight so confirmatory of their views as to the greater density of cast iron in the molten than in the solid state , that it seems necessary here to present the true explanation of the facts , which , so far as they are here relevant , may be best given briefly in the words of these authors :\#151 ; " When a ladle of molten iron is drawn from the furnace and allowed Mr. R. Mallet on the alleged Expansion to stand at rest , the thin coat of scoria or molten oxide which forms on the surface of the metal is seen , as fast as it forms at the circumference of the ladle , to be swept by active convergent currents towards the centre , where it accumulates in a patch .
As the fluid metal parts with some of its heat and the ladle gets hot by absorbing it , this remarkable surface-disturbance becomes less energetic .
" This arises from " the expansion of that portion of the molten mass which is in contact with the comparatively cool sides of the ladle , which sides act as the chief agent in dispersing the heat of the melted metal ; careful observation will show that the motion in question is the result of an upward current of the metal around the circumference of the ladle , as indicated by the arrows A , B , C in the accompanying sectional drawing of the ladle " [ fig. 7 ; the figure in Fig* 7 .
the margin is essentially the same as that of these authors ] .
" The upward current of the metal can be seen at the rim of the ladle , where it is deflected into the convergent horizontal direction , and where it presents an elevatory appearance , as shown in the figure .
It is difficult to assign to this any cause but that of expansion and consequent reduction of specific gravity of the fluid metal in contact with the sides of the pot , as , according to the generally entertained idea , the surface-currents above referred to would be in the contrary direction to that which they invariably take , i. e. they would diverge from the centre instead of converging to it .
" The facts , so far as they are above described , are generally correct , but the explanation given is not the true one .
The currents observable for some time after a large ladle ( say , holding 10 tons ) is first filled with molten iron are not produced by difference of temperature in different parts of the mass , but in the following way :\#151 ; Such a ladle is of wrought iron , about half an inch in thickness ; and to preserve this tolerably cool , even for several hours , it is lined with a coating of earthy material daubed upon the interior in a tough and plastic state , from an inch to an inch and a half in thickness , and dried within it .
The lining material consists of plastic clay , with a proportion of siliceous sand beaten up together with horsedung , chaff , plasterer 's cow-hair , or other fibrous material , conferring toughness upon the mass when soft and porosity when dry .
This material , after drying at a temperature averaging 500 ' to 700 ' Pahr .
, on being exposed to contact with the molten cast iron , exhales torrents of gas and vapour , which pass upwards through the of various Substances on Solidification .
molten mass and determine the direction of its currents ; and it will be obvious , on inspecting the figure , that these currents will be most powerful round the outer circumference of the mass , where each unit of its top surface has a larger proportion of lining in proximity to it than at the central parts of the mass , where downward currents are the necessary consequence of those produced upwards at the circumference .
The organic matters mixed with the lining are carbonized , and give forth the elements of water as well as nitrogen .
The clay , which is a hydrous silicate of various earthy bases , gives forth its water and some of the oxygen of the peroxide of iron which most clays contain .
More or less carbonate of lime is almost always interspersed , and this gives forth carbonic acid and water .
The gases thus streamed forth act mechanically by their ascent and also chemically upon molten iron , the water being decomposed , oxidizing portions of the iron and forming scoriae , which is again more or less reduced by contact of the hydrogen and nitrogen when the latter is present .
These rapid combinations and decompositions are no doubt the main cause of those singular vermicular startings referred to by Messrs. Nasmyth and Carpenter , which are familiar to every iron-founder , but which are entirely distinct from the ascending and descending currents due to the ascent of the evolved gases .
That this is the true explanation is supported by the following facts:\#151 ; 1 .
After a large ladle has stood full of molten metal for some hours , and time has been given thus for the whole of the gaseous contents of the lining to be driven off , the ascending and descending currents cease to be perceptible , and if any currents at all can be discerned they are in the opposite directions .
2 .
If , after this , such a ladle be emptied of its contents , the lining remaining untouched and only coated with a thin shell of adherent cast iron [ and oxides and silicates of iron ] , and the ladle being again filled with molten iron , no such currents as at first are produced in the molten mass , the lining having been previously exhausted of its gases and vapours .
That the currents described by Messrs. Nasmyth and Carpenter are not due to dissipation of heat from the mass through the sides of the ladle is evident from the following considerations :\#151 ; A 10-ton ladle , which is about 4^ feet by 3 feet in depth , loses heat so slowly that after standing for six hours the molten metal is still fluid enough to make castings .
Let us suppose it filled into the ladle at a temperature of 2800 ' to 2900 ' Fahr. , and that after six hours it is still 200 ' above the temperature of solidification of cast iron , or at 2600 ' .
The molten mass has thus lost 300 ' of heat in 360 minutes , or *0138 of a degree per second .
We may assume this at any instant as representing the difference in temperature between two vertical columns , one at the centre and the other at the circumference of the molten mass .
The linear dilatation of cast iron for one degree of Fahrenheit being O'OOOOlll , as deduced from its total cubic dilatation between 60 ' Fahr. and the temperature of fusion at which it was poured into the cone , as given in this Mr. R. Mallet on the alleged Expansion paper , and assuming the depth of the colder of these columns , whether that be at the circumference or not , to be , as stated , 36 " , that of the hotter column will be 36-0000005514 , and the difference between these two measures the force which alone can produce circulating currents in the mass by differences of temperature due only to cooling .
This is equally true whether it be the colder column that is dilated , as supposed by Messrs. Nasmyth and Carpenter , or the hotter one , as is the fact .
And if we consider the viscidity of molten cast iron , it is perfectly obvious that the circulating currents referred to by Messrs. Nasmyth and Carpenter cannot be due to so insignificant a cause .
Want of attention , or careless interpretation of the many and somewhat complicated conditions thus seen to be involved in the cooling of a solid by dissipation of its heat from its exterior , has caused many serious misapprehensions on the part of experimenters as to the supposed expansion of metals in volume when consolidating .
Thus , even in the case of bismuth , it has been supposed a conclusive proof of its expansion that a mass cooling in an open crucible exudes from its interior upon its top surface cauliflower-like excrescences ; but although the author does not here deny or affirm any thing as to expansion being a fact in the case of bismuth , it is nevertheless obvious that such excrescences might arise merely from the grip of the crucible itself , or even of the exterior portions of the metal already solidified contracting upon and so squeezing out portions of the still liquid interior .
It is stated on good authority that a distinguished artillery officer , in former years at the head of the Laboratory , Woolwich Arsenal , satisfied himself of the reality of the expansion of cast iron in consolidating by the following experiment An elongated projectile was cast , with its axis vertical , in a very thick and massive mould of cast iron , the mould being cold or nearly so ; the molten metal was introduced through a narrow aperture applied at the base of the projectile , as in fig. 8 , the neck or " gate " being knocked off the instant the mould was filled .
As the cooling rapidly proceeded , portions of the still fluid metal were forced out at the place where this neck was detached ; and the conclusion was come to that the exterior being already solidified such excrescence could only arise from expansion of the contained liquid metal as it solidified in succession .
What really did take place , and is the true explanation Fig. 8 .
of various Substances on Solidification .
t$\#163 ; ie facts , is , that when a very thick iron mould of this sort is suddenly Vjwed by pouring molten iron into its interior , as the heat abstracted Iflfra the latter can only pass into the material of the mould at a rate dermined by its conductivity , so the interior part rapidly becomes miiid to a temperature enbrmously higher than the exterior portions , tirbh for a time remain almost cold .
The expanded interior walls of hii mould push inwards as towards the points of least resistance , and so * r ally diminish the capacity of the mould for a time , the inner surfaces of hich press upon the consolidating crust of metal within it , and so seze out in part its liquid contents , just as water might be squeezed fom an india-rubber bottle* .
o seemed desirable to obtain some experimental results in reference to th objects of this communication with lead .
It has never , so far as the amor is aware , been even suggested that this metal expands in consoli-dsng .
Its coefficient of dilatation by heat is enormously greater than tih ; of cast iron , being , according to the determination of Lavoisier and Jj lace , between 32 ' and 212 ' Fahr. =0*0000474 of its volume for one three Fahr. ; so that , taking its fusing-point at 617 ' '(Eudberg ) , and atuning the coefficient constant for the entire range from 60 ' to 617 ' sfiich is much below the truth ) , its dilatation when in fusion would he 0*0264 of its volume , and the specific gravity of lead at 60'= 11*36 ; Hfc of liquid lead must be below 11*07 .
Indeed this enormous amount ^dilatation is impressed upon any observer who sees the rate at which al lead in casting a common bullet sinks into the neck of the mould , the comparatively large cavity which always exists in the ball when n in two .
From its low temperature of fusion and the suddenness h which lead passes from the solid to the liquid state without mtphase of intermediate viscidity , and only a brief one of crystalline botleness , and the facility with which its surface can be kept free from d'ss or oxide , this metal presents a " crucial " example for experiment Preference to our subject .
The following experiments , by the kind permission of Messrs. Pontifex M Wood , London , were made at their works :\#151 ; Lst .
Upon the surface of a large pot of melted lead , the temperature ' I * The enormous disparity in temperature between the interior and the exterior cokes of a very thick spherical or cylindrical iron mould is only partially shown by ht 's expression , log p the distribution of temperature along a bar heated at one extremity ; for in the case 1 e before us , as in that of all forces radiating from a centre , the temperature at any fen point , and quite independent of any question of conductivity , varies inversely as square of its distance from the centre .
If a unit in volume situated in any point ot Radius lose as much heat as will lower its temperature 1 ' , that heat , when trans- ) lotted to double the distance along the radius , can only raise the temperature of a unit Volume there by 0*25 of 1 ' having been diffused amongst four units of volume .
( See -ot , * Traits de Physique , ' and Mallet/ Trans. R. I. A. , ' 1856 .
) 230 Mr. R. Mallet on .
the alleged Expansion of which was estimated at from 750 ' to 880 ' Fahr. , the half of a large pig of newly smelted lead , being a semicylindrical bar of about 5 " x 2i " and about 18 " long , was gently laid down horizontally ; it immediately sank to the bottom and there remained .
When about half its volume was melted away , the unfused portion was drawn up to the surface and let go , when it at once sunk to the bottom again .
2nd .
A ball of such lead was cast , weighing 17| lbs. , diameter about 4|- " ; this was put into an empty hand-ladle , which was gently placed upon the surface of the pot of melted lead ; the ladle was depressed sufficiently to fill with lead , and being left free was carried to the bottom of the pot with sufficient impetus to produce a sensible blow of the exterior of the ladle upon the bottom of the pot .
3rd .
A flat circular disk of about 1*25 inch in thickness , being laid gently upon the surface , after a moment 's hesitation slowly went to the bottom .
Another disk of 6 " diameter , by rather less than an inch in thickness , remained a few seconds longer on the surface and then sunk to the bottom ; both disks , while they floated , had their top surfaces but very slightly elevated above that of the liquid lead .
One of the disks being gently lowered into the liquid lead vertically and edgeways , at once went to the bottom .
4th .
Two disks , each 6 " diameter , the one 0*57 inch and the other 0*4 inch in thickness , being gently laid flat upon the surface of the molten lead , floated , and with an emergent portion sensibly greater than that of the disks in experiments 2 and 3 , and remained floating until about 1*25 of the radius had been melted away all round , when they slowly sunk in the liquid , as was proved by the slow disappearance of the slender iron wire cast into the middle of the disk for the purpose of lowering .
The thinner of these two disks floated rather longer than the thicker .
5th .
A plate of sheet or laminated lead , clean from the rolling-mill , of about 0"*5 in thickness and about 10 " square , being gently placed flat on the surface of the liquid lead , floated , its top surface being nearly level with that of the liquid .
After about ten seconds a piece was melted off from one of the edges , when the plate canted in the opposite direction and sunk .
6th .
Plates of about 0"*18 thick floated much in the same manner as the preceding .
The temperature of the solid lead employed was in all cases about 70 ' Fahr. It follows from these experiments that , as in the case of cast iron , the solid does not float upon the liquid lead through buoyancy , that , on the contrary , the negative buoyancy is very marked , and that the repellent force , whatever be its nature , by which flotation is produced is dependent upon the effective surface as compared with the volume of the solid .
They present also a corroboration of the view that the repellent force upon the unit of effective surface is greater as the difference of tempera* ture between the solid and liquid metal is so .
of various Substances on Solidification .
; I proceed to some remarks upon the experiments referred to at the \lt ; mmencement of this paper , and quoted by Messrs. Nasmyth and Car-inter , as to the floating of pieces of solidified iron furnace-slag upon the me slag when in the liquid state .
It is a fact that blast-furnace slags oled below the point at which they become rigid do very generally \gt ; at upon the same slag in its molten state .
It is equally true that the tsic silicates which constitute the chief part of terrestrial volcanic lavas ?
at upon the surface of these when molten .
But these admissions do )t suffice in any degree to support the conclusion deduced by Messrs , asmyth'and Carpenter , that basic silicates , whether as furnace-slags or vas , are denser in the molten than in the solidified state , nor that these ; odies in the act of solidification expand in volume or decrease in density i any manner , irrespective of the formation or enlargement of cavities cgas-bubbles within them .
The experiments of the author upon the )tal contraction of iron furnace-slags for the entire range of tempera-are between that of the blast-furnace and the atmosphere , made at the { arrow Iron-Works , and fully described in the author 's paper on " The Mature and Origin of'Volcanic Heat and Energy , " printed in Phil , f 'rans .
for 1873 , leave no doubt as to the following facts :\#151 ; I 1st .
That the density of such slags at 53 ' Eahr .
is to their density when molten and at the temperature of the blast-furnace as 1000 : 933 , r , taken at the melting-point of slag , as 1000 to 983\#151 ; molten slag being hus very much less dense than the same when solidified . !
2ndly .
That no expansion in volume whatever occurs in such slags at \gt ; r near the instant of solidification .
The experiments of the author above referred to were made by filling 'ast-iron slightly conical moulds with the slag run direct from the blast-umace , and permitted to consolidate and cool therein , by which perfectly solid slightly conical blocks were obtained .
From the method employed , and the very large scale upon which these experiments were conducted , it is impossible that any expansion in volume at or near the point of consolidation , if even of a very minute amount , could have occurred and yet have escaped notice* .
It is only necessary for the author here to point out that the floating of crusts of slag or lava is not due to the cause assigned by Messrs. Nasmyth and Carpenter ; nor is it his intention to enter at any length into what are the causes of such floating when it occurs .
The following remarks , however , may be made :\#151 ; It is impossible to obtain a moderate-sized fragment of solidified slag or lava free from air-bubbles , and from involved or superficial cavities , which tend to float the mass when thrown upon its own material in the melted state .
Those who have attentively watched large volumes of slag issuing from the blast-furnace are aware that it comes forth carrying with it a large * [ For the proofs of which in detail the author begs to refer to his paper at length , Phil. Trails .
1873 .
] S VOL. XXIII .
T Mr. R. Mallet on the alleged Expansion volume of gaseous matter minutely diffused , which is pretty readily sepa rated , and is characterized by a white vaporous cloud floating thinly ove ' the issuing stream ; if the slag be cooled rapidly , the gaseous or vapori zable bodies present become confined and render the mass vesicular , whit if cooled more slowly , and with a free surface for the escape of these the mass solidifies more solidly , often as solidly as a block of granite* .
Independently of the buoyancy that is produced by the vesicularity \#169 ; rapidly cooled slags , it is highly probable that relatively cold and solif slag , whose buoyancy is negative , may yet float on molten slag , whose \gt ; density is less than its own , in virtue of that same repellent force which as we have seen , acts under like conditions in the case of metals .
With respect to acid silicates , or slags analogous to glass ( which , how ever , are not referred to by Messrs. Nasmyth and Carpenter ) , the autho again refers to the results given in his paper ( Phil. Trans , 1873 ) .
These and indeed the circumstances attending the production and destructioi of the well-known " Rupert 's drops , " incontestably prove that these sili cates also are less dense in the molten than in the solid state , and tha they contract violently at or near the instant of consolidation .
j The author has more than once heard the opinion expressed by those engaged about blast-furnaces , that their slags do expand in consolidating based upon a misinterpretation of the following frequently occurring cir cumstance:\#151 ; When the large parallelopipeds of slag ( 5 to 6 feet square b ; 2 to 3 feet thick ) are stripped from the iron square frame which formei their edges , and are being removed upon the iron waggons on which the ; are cast , and still , as often happens , in a very hot state , or even with ; still liquid or viscous interior , though rigid externally , it occasional !
happens that such a block bursts asunder , and with a suddenness which is sufficient sometimes to scatter dangerously some of the liquid interior ; or if the fracture be not so sudden , and the interior be in a viscous condition , the latter may continue for a considerable time to slowly exude in fantastic shapes from any aperture of escape left free to it ( see fig. 9 ) .
These facts have been supposed to indicate that the interior of the mass expands in consolidating .
It is scarcely Fig. 9 .
a. Tail of yet viscid slag extruding from interior here , however , to eBP\#174 ; necessary All this may be well observed in the millions of blocks of slag from the Clevels\#174 ; blast-furnaces of which the vast embankments and breakwater at the estuary of W IJiver Tees have been formed .
of various Substances on Solidification .
nit any detail to prove that the phenomena are due to the contrac-t L of the already solidified exterior upon the unyielding interior of the ms ; the former becoming fractured by its own grip , and its material nog highly elastic , often yields with apparently explosive violence like iddenly broken spring ( see fig. 9 , which shows the appearance of one olhese fractured and slowly exuding blocks ) .
The following remarks may be made , in addition to those preceding , ncontravention of the supposed expansion of slags or lavas in consoli-I ing .
It is well known that masses of mud when dried by the sun 3k , the fissures penetrating nearly perpendicular from the surface and -narating into more or less symmetrical prisms .
Blocks of starch after diccation present similar phenomena , which are also frequently seen simplified by the uppermost beds of argillaceous limestone ( or calp ) of Hand when first laid bare from its detrital covering .
In all these cases i re can be no doubt that the phenomena are due to the shrinkage of fe mass in drying .
But shrinkage or contraction by cooling and consoli-cion ought to present us with like results ; and these we see actually i infest in the splitting-up of basalt into columnar prisms whose long na are always found perpendicular to the surface by which the heat of t } mass was dissipated .
Such columnar separation is not confined to tsalt ; instances of it are abundant in lavas of every age , \#171 ; the surfaces \lt ; the prisms in these being sometimes straight , sometimes curved , though much remains yet to be investigated before all the circumstances sending the splitting-up of masses of basalt or lava can be said to be F risk ymderstood , y\#174 ; t enough is already known and clearly explained to ike it certain that it is due to contraction of these materials as they ol , and that this form of splitting-up is wholly incompatible with that any fissurmg that could arise from the refrigeration of a mass the lume of every part of which expanded in consolidating .
] As m what precedes the hypothesis upon which the lunar volcanic } '* M'essrs* Nasmyth and Carpenter rests is proved to be without a ion , it seems needless to enlarge upon the incongruities and con-admtions which the theory itself presents when fairly applied to such , a\#174 ; we have of the volcanic features of the moon , or still more e *2 ' .i ltmust be were 'lt true\gt ; to those of our earth , [ assuming iemi\#153 ; ienalS 'lOUr 6arth and satellite analogous in their physical and teir wnlrTrtl\lt ; !S~an assumPtion made by these authors throughout In n l ' ^0Ugh , withlout any attempt to support it by proof ] .
beralitvCf\gt ; U / T*8 ^1S paPer ' the author has to express his thanks for the tU8 " quiry of " .***"* \#187 ; Wacter and with-^ \#171 ; appUcabUity- He \#171 ; 1*\gt ; to thank Mr. E. iso has to a 'V^ .
^'r Persona^ aid and cooperation .
He V0L"x\#171 ; .eXpreS8 ^ the valuable a..i\#171 ; tanee so readily 234 Dr. W. B. Carpenter on Nature of the [ peb 4 afforded him by Colonel Milward , It .
A. ( Superintendent of the Laboratories , Royal Arsenal , Woolwich ) , and to Mr. Davison , his chief assistant .
Like thanks he wishes to return also to Messrs. Pontifex and " Wood .
Lastly , he wishes to record the valuable aid he has received in making these experiments and calculations from his assistant , Mr. " W " Worby Beaumont .
February 4 , 1875 .
JOSEPH DALTON HOOKER , C.B. , President , in the Chair .
The Presents received were laid on the table , and thanks ordered for them .
The following Papers were read I. " Remarks on Professor Wyville Thomson 's Preliminary Notes on the Nature of the Sea-bottom procured by the Soundings of H.M.S. 'Challenger .
' ' ' By William B. Carpenter , M.D. , LL. D. , E.R.S. Received Dec. 29 , 1874 .
The extreme interest of two of the questions started and partly discussed in Professor Wyville Thomson 's communication will be deemed , I trust , a sufficient reason for my offering such contributions as my own experience furnishes towards their solution .
The first of these questions is , whether the by the accumulation of whose shells the Globigerina-ooze is being formed on the deep-sea bottom , live and multiply on that bottom , or pass their whole lives in the superjacent water ( especially in its upper stratum ) , only subsiding to the bottom when dead .
Having previously held the former opinion , Prof. Wyville Thomson states that he has now been led to adopt the latter , by the results of Mr. Murray 's explorations of the surface and sub-surface waters with the tow-net\#151 ; which results concur with the previous observations of MulleL Haeckel , Major Owen , and others , in showing that Globigerince , in common with many other Foraminifera , have a pelagic habitat ; while the close relation which they further indicate between the surface-fauna of any particular locality and the materials of the organic deposit at the bottom , appears to Prof. Wyville Thomson to warrant the conclusion that the latter is altogether derived from the former .
Now without in the least degree calling in question the correctness of these observations , I venture to submit , first , that they bear a different interpretation ; and second , that this interpretation is required by other facts , of which no account seems to have been taken by Prof. Wyville Thomson and his coadjutor .
In this , as in many other instances , I '\lt ; \#166 ;

