]\gt ; PHILOSOPHICAL TRANSACTIONS .
I. On the ditions of } ) of Theories of Cosmogony .
G. H. DARWIN , .D .
, F.R.S. , Fellow of Trinity College and in the Unirersity of mbridge .
Received July 12 , --Rcad November 15 , 1888 . .
LOCKYER writes in his .
paper on as follows:\mdash ; ' ' The brighter lines in spiral nebulae , and in those in which a rotation has been set up , are in all probability due to streams of meteories with irregular motions out of the main streams , in which the collisions would be almost nil .
It has already been suggested by Professor G. DARWIN the seous hypothesis\mdash ; that in such nebulae the great nlass of the gas is non-luminous , the luminosity an evidence of condensation along lines of low yelocity , according to a well known hydrodynamical law .
From this point of view , the visible nebula may be regarded as a luminous diagram ofits own stream-lines .
' " " The whole of Mr. LOCKYER'S paper , and especially this passage in it , leads me to make a suggestion for the reconciliation of two apparently divergent theories of the origin of anetary systems .
The nebular hypothesis depends essentially on the idea that the primitive nebula is a rotating of fluid , which at successive epochs becomes unstable from excess of rotation , and sheds a ring from the equatorial region .
The researches of ROCHE ( apparently but little known in this country ) have imparted to this theory a precision which was wanting in LAPLACE'S original exposition , and have rendered the explanation of the origin of the planets more perfect .
But notwithstanding the high probability that some of the kind is true , the acceptance of the nebular hypothesis presents great difficulties .
Sir WILLrAM long ago expressed to roe his that the most probnble origin of the planets was through a gradual accretion of meteoric matter , * ' Nature , ' Nov. 17 , 1887 .
The paper itself is in ' Boy .
Soc. , 188 .
'Montpellier , Acad. Sci. NIem .
' [ The remarkable photograph of the nebula in Andromeda , to the onomical Society by .
Isaac Roberts on December 6 , 1888 , affords something like a of the of the othesis. .
H. J ) .
December 19 , 1888 .
] G. H. D-kRWlN ON TIf ICAL OF the re , rehes of Mr. LOCKYER afford actual evidence in your of the abundance of lnebeorites in space .
But the very essence of the nebular hypot.hesis is conception of fluid pressure , since without it of a of equilibrium becomes inapplicable .
Now , at first sight , the meteoric condition of matter seems absolutely inconsistent with a fluid exercised by one pal't of the system on another .
We thus seem driven either to the absolute rejection of the nebular hypothesis , or to deny that the meteoric condition was the immediate antecedent of the Sun and Planets .
M. FAYE has taken the former course , and accepts as a necessary sequence the formulation of a succession of events quite different from of the nebnlar ' I cannot myself his theory is an improvement on that of LAPLACE , except in regard to the adopof meteorites , for he has lost the concel ) tion of the figure of equilibrium of a rotating mass of fluid .
The object of this paper is to point out that by a certain interpretation of the meteoric theory we obtain a reconciliation of these two orders of ideas , and may that the of stellar and planetary systems is meteoric , whilst the conception of fluid to the kinetic theory of gases , fluid pressure is the average result of the of molecules .
If imagine the molecules magnified until of the size of orites , their impacts will still , on a scale , give a quasi-fluid pressure .
I then that the fluid pressure essential to the nebular hypothesis is , in fact , the resultant of countless impacts of meteorites .
The of hydrodynamics could hardly be attacked with success , if we were forced to start from the beginning and to consider the cannonade of molecules .
But when once satisfied thac the kinetic theory will give us a gas , which , in a space containing some millions of molecules , obeys all the laws of an ideal non-molecular filling all space , we may put the molecules out of sight and treat the gas as a In the same way , the difficulty of tracing impacts of meteol'ites in detail is insuperable ; but , if we can find that such impacts give rise to a quasi-fluid pressure on a large scale , we may be able to trace out many results by treating an ideal plenum .
LAPLACE'S hypothesis implies such a plenum , and it is here maintained that this lenu m is merely the idealisation of the impacts of meteorites .
As a suggestion this view is worth but little , for its acceptance or rejection must turn entirely on numerical values , which can only be obtained by the consideration of some actual system .
It is obvious that the solar system is the only one about which we have sufficient knowledge to afford a basis for discussion .
This paper is ccordingly devoted to a consideration of the mechanics of a swarm of meteorites , with special numerical application to the solar system .
The investigation iled a considerable amount of mathematical analysis ; * l'Origine du Moltde , ' Paris , , 1884 ; 'Annuaire pour l'an 1885 , Bureau des Lon A SWARM OF TEORl T , AND ON COSIIOGONY .
there is , however , no analysis in and 2 .
The reader who only wishes to know the arguments ] results , without a consideration of the mathematical details , is therefore recommended , after reading SS1 and 2 , to pass on to the Summary .
S 1 .
On the Effecfive Elasticity of Meteorites Collision .
When two meteoric stones meet with planetary velocity , the stress between them during impact must generally be such that the limits of true elasticity are exceeded ; and it may be urged that a kinetic theory is unless the colliding partides are highly elastic .
It may , however , I , be shown that the very greatness of the velocities will impart what amounts to an elasticity of a high order of perfection .
It appears , a priori , probable that , two meteorites clash , a of the solid matter of each is volatilised , and Mr. LOCKYER considers the spectroscopic evidence conclusive that it is so .
There is , no doubt , enough energy liberated on impact to volatilise the whole of both bodies , but only a small portion of each stone will undergo this change .
A rough numerical example will sllow the kind of quantities with which we are here dealing .
It will appear hereafter that the mean velocity of a meteorite may be at the least about 5 kilometres a second ; and , accordingly , the relative velocity of a would then be about 7 kilometres a second .
Hence , if two stones , a kilogramme , move each with a velocity of kilometres per second directly ; one another , the energ.y liberated at the moment of impact is or ergs .
Now JOULE'S is ergs ; hence , the energy libel.atod is about 3 million calories .
It is quite uncertain how much of each stone would be volatilised ; but , if it were 3 grammes , there would be a million calories of energy applied to each gramme .
The temperature of iron is about 1500 degrees rade , and the mean specific heat of iron may be about Hence , about 300 calories are reqnir to raise a graYnme of iron from absolute zero to melting point .
I do not know the latent heat of the melting of iron , but for platinum it is 27 , and the latent heat of tion of mercury is 6 Hence , about 400 or 500 calories suffice to a of iron from absolute zero to volatilisation .
It is clear , then , that there is enercry enongl ] , not only to volatilise the iron , but also to render the gas ; and the same would be true if the mass operated on by the energy were 30 grammes instead of 3 .
It must necessarily be obscure as to how a small mass of solid matter take up a very large amount of energy in a small fraction of a second , but ) ectroscopic evidence seems to show that it does so ; and , if so , we have what is tually a explosive introduced the two stones .
be the velocity of mean sqnare , is the square of the mftll qunre of ' ' PROFESSO , .
H. DARWIN ON THE MECHANICAL CONDITIONS OF In a direct collision each stone is probably shattered into fragments , like the splashes of lead when a bullet hits iron target .
But direct collision must be comparatively rare event .
In glancing collisions the velocity of neither body is wholly arrested , the concentration of energy is not so enormous ( although probably still cien t to effect volatilisation ) , and , since the stones rub past one more time is allowed for the round the point of contact to take up the energy ; thus , the whole process of collision is much more intelligible .
The nearest terrestrial analogy is when a cannon-ball bounds ofl the sea .
In glancing collisions fracture will not be very frequent .
these uments , it is probable that , when two meteorites meet , they attain an effective sticity of a high order of perfection ; but there is , of course , some loss of at ejlch collision .
[ It must , however , be admitted that on collision the flection of path is rarely through a very large .
But a succession of glancing collisions would be able of reyersing the path ; and , thus , the kinetic theory of meteorites may be taken as not differing materially from that of gases .
* ] Perhaps the most serious difficulty in the whole theory arises from the fractures which must often occur .
If they happen with great frequency , it would seem as if the whole swarm of meteorites would rade into dust .
We know , however , that meteorites of considerable size the Earth ; and , unless Mr. LOCKYER has eted the spectroscopic evidence , the nebulae do now consist of meteorites .
Hence , it would seem as if fracture was not of very fiequent occurrence .
It is easy to see that , if two bodies meet with a giyen velocity , the chance of fracture is much if they are large , and it is possible that the process of breaking up will go on only until a certain size , dependent on tlJe velocity of agitation , is reached , and will then become comparatively unimportant .
When the volatilised gases cool , they will condense into a metallic rain , and this may fuse with old meteorites whose surfaces are molten .
A meteorite in that condition will certainly also pick up rlust .
Thus , there are processes in action , to counteract subdivision by fracture and volatilisation .
The meam size of meteorites probably depends on the ance between these opposite tendencies .
If this is so , tbere will be some some fnsions , but the mean mass will change very with the kinetic energy of agitation .
This view is , at any rate , adopted in ths paper as a working hypothesis .
It was not , however , possible to take account of fracture and fusion in the mathematical , but the meteorites treated as being of invariable mass . .
On the Velocity of Meteorites , on its Secular The velocity with which the meteorites move is derived from their fall from a great distance a centre of egation .
In other words , the potential energy of their mutual actio when widely dispersed becomes , at least partially , *Added Nov. 16 , 1888 .
A SWARII OF ] into kinetic energy .
When the condensation of a swarm isjust beginning , the mass of the aggregation towards which the meteorites fall is small ; alJd , thus , the new bodies arrive at the aggregation with small velocity .
Hence , initially , the kinetic energy is small , and the of the sphere within which hydrostatic ideas are ( if aIlywhere ) applicable is also small .
As more and more meteorites fall in that volume is enlarged , and the velocity with which they reach the tion is increased .
Fioally , the supply of meteorites in that part of space begins to fail , and the imperfect elasticity of the ] liding bodies brings about a gradual contraction of the swarm .
I do not now attempt to trace the whole history of a swarm , but the object of the paper is to examine its mechanical condition at an epoch when the supply of meteorites from outside has stopped , and when the velocities of itation and distribution of meteorites in space have arranged themselves into a sub-permanent condition , only affected by secular .
This examinatior ] will enable us to understand , at least roughly , the secular in the velocity and in the distribution of the meteorites as the swarm contracts , and will throw light on other questions .
S3 .
Fomuloe for Mean Square of Vdocity , ] , Frec th , betne Collisions .
We have to investigate whether , when the solar system consisted of a swarm of meteorites , the velocities and encounters could have been such that the mechanics of the system can be treated as subject to the laws of hydrodynamics .
The formulae which form basis of this discussion will now be considered .
For the sake of sirnplicity , the meteorites will , in the first instance , be treated as spheres of uniform size .
The sum of the masses of the meteorites is equal to that of the Sun , for the planets only contribute a negligible mass .
If be the Sun 's mass , and that of a } eteorite , their number is If , at each encounter between two meteorites , there were no loss of energy , the sum of the kinetic energies of sll the meteorites would be equal to the potential energy lost in concentration of the swarm from a condition of infinite dispersion , until it possessed its actual arrangement .
In such a computation the rotational of the system is negligible .
Suppose the SUD'S mass to be concentrated from infinite disp ersion until it is arranged in the form of a homogeneous sphere of radius and density .
Then let bhe sphere be cut up into as many equal spaces as there are nleteorites , let matter in each space be concentrated into a teorite .
When the number of meteorites is large , the potential energy lost in the first process is very great } ) with that lost in the subsequent partial conden sation lneteolite energy lost in the condensation is negligible .
* It depends , in fact , on the square uf tho atio of to of of the equal spaces .
SSOR Cr .
H. ON CHANICAL CONDITTONS If be the attractional constant , the lost energy of condensation is well known to be .
But on the hypothesis that there is no loss of energy at each en counter , this must be equal to the sum of the kinetic energies of all the meteorites .
If , therefore , be the mean square of velocity of a meteorite , we must have so that But homogeneity of density and uniformity of kinetic energy of agitation are impossible ; for the meteor-swarm must be much condensed towards its centre , so that we have largely underestimated the lost potential energy of the system .
Also , velocity of agitation must towards the outside , or else the swarm would extend to finity .
Besides this , the partial of molar into molecular energy , which must take pl on each encounter , has been neglected .
We see below reason for believing that a large central volume the mean square of velocity of agitation is lJearly uniform , and that outside of this region it falls off .
, then , that is the and the radius of that portion of the swarm in which the square of velocity of agitation is unifo1m ; let be that square of velocity , and let it be reference to the of I at distance , so that , ( 1 ) where is a coefficient for which a numerical ) will ) found The square of velocity of tion outside of the radius is to be denoted by and sequent ition will be necessary to uate v in terms of If we denote by the Earth 's distance from the Sun , and by Earth 's velocity in its orbit , we . . . . . . . . . . .
( 2 ) Whence , . . . . . . . .
( 3 ) If in any distribution of ites r is sum of the masses of all the meteorites in unit volume , or the density of the swal at any point , and if be that distance which is called the kinetic theory of " " the mean distance between neighbouringr molecules we have ( 4 ) Now , the mean den sity of that of the swarm in which the kinetic energy of agitation consta1lt being , we have , . . . . . . . .
( 5 ) and .
( 6 ) A SWARM OF rEORfTES , AND ON THEORTES OF COSlfOGONY .
Suppose that is " " the radius of the sphere of acbion\ldquo ; of a meteorite , so that when two of them approach so that the distance between their centres is there is a collision .
Let and be the mean free path and mean interval between ions .
Then , since the mean velocity is , we accordin to the kinetic theory of gases , * . . . . . . .
( 7 ) Then , on from ( 4 ) , ( 5 ) , and , we have ( 8 ) Now let , ( 9 ) and we have . .
( 10 ) We now proceed to , and also the centim e second system of units .
The Sun 's mass may be taken as 315,511 times that of the E.:rrth , and the Earth as grammes ; hence grammes .
The attractional constant and the 's mean distance from the Sun are .
, 'Kinetische Theory der Gase .
' and elsewhere I alJy use EVERETI'S ' Units and Physical onstants .
' PllO SSOR G. ON TII IIECHANICAL CONDITIONS Witb values cm .
per sec. , 800,600 cm . .
( 11 ) sec. The dimensions of nd are not length and time ; but , if teorites of 1 gramme mass , with sphere of action 1 centimetre , and " " velocity of mean square of agitation to the 's velocity in its orbit , have density of distribution equal to one-third of the mean density of the sphere , then will be the mean free path and time , as tated in centimetres and seconds .
We may thus regard as a length and time , provided care be tadeen in the subsequent use of the symbols to adhere to the c.g.s. system of units .
S4 .
the of a Gas Uniform Temperature Concen tSpherical der its own vitatio It is assumed provisionally the conditions are fied which permit us to regard the swarm of meteorites a , s a quasi-gas , subject to the laws of hydrostatics .
The solution of this problem , then , becomes a necessary preliminary to the discus- sion of the kinetic theory of teorites .
The equilibrium of a gas under its own gravitation has been discussed by Professor RITTER in one of his series of papers on gaseous planets .
The intrinsic interest of the problem renders an independent solution yaluable .
Suppose , then , that a mass of is enclosed in a spherical envelope of radius , and is in equilibrium in concentric spherical ]ayers .
Let the mean square of the velocity of itation of the aseous molecules , be defined by reference to the potential of the mass at the radius , so that where is a numerical coefficient , and is the attractional constant .
Let be the pressure and density of the gas at radius , and the modulus of elasticity , so that * ' ' Untersuchungen iiber die Hohe der Atmosphare und die Constitubion miger Weltkorper , \ldquo ; ' Annnlen ' ( New Series ) , vol. 16 , 1882 , p. 166 .
A very elegant solution of pa rt of my oblem 1 also been given by .
G. W. in the ' Annals of Mathcmatics , ' vol. 4 , No. 1 , p. 19 obruary , .
Mr. only reached my hands my own calculations been completed , I to my ow less method .
bas obviousJy not seen pers .
A SWARM OF METEORITES , AND THEORIES OF COSMOGONY .
Then the equation for the hydrostatic equilibrium of the gas is .
( 12 ) It is obvious that \mdash ; \mdash ; equal to the whole mass enclosed inside radius and this relation will hold however the equation be transforrned , provided we do not multiply the equation by any factor .
In consequence of the relation between and this may be written If be the mean density of the mass , we have Hence , we may write the equation ( 12 ) in the form Now , let and the equation becomes .
( 13 ) By erentiation we obtain the equation It is obvious from ( 13 ) that is he mass enclosed inside radius and therefore is equal to unity when A general analytical solution of ( 14 ) does not seem to be attainable , and must be had to numerical processes .
Although this is an equation of second degree , and its general solution must involve two arbitrary constants , we shall see ( as pointed out by M. RITTER ) that the general solution , as applicable to our problem , may be deduced from one single numerical solution .
M. proceeds a graphical method , which he has worked with surprising accuracy .
I shadl therefore adopt an analytical and method , which , ) , is suscept i greater accuracy .
DCCCLX X. tever be the al.rangement of the gas , the density at the centre must have some value .
I therefore start with a central density , corresponding to the value of sc that .
( 15 ) For the sake of brevity the suffixes 1 will be now omitted from the various symbols , to be reaffixed later when they ar6 reqnired .
At the centre , where is , \amp ; c. , all zero , and we put Let , and let us assume . .
Now , the differential tion ( to be satisfied is But and by we . .
By equating coefficients in these two series , I find ' \amp ; c. , and ; whence , by extolt.n , In .
RtTTER'S paper , already referred to , he takes a cel.tain function as equal to the radius is unity .
lnction ?
in my notation to A SWARM OF METEO R , AND ON OF COSMOGONY .
or .
It follows , therefore , that RITTER takes the value of .
But he intends the central density to be 100 times the sul.face density ; hence , to the same solution , we must have equal to 100 times .
Therefore , his value of would be or Now , as I want to make a comparison between my solution and his , I start with this value of .
The only object attained by the choice of this particular value that the two solutions become easily comparable .
It will be seen below the alue of does not make the central density exactly 100 times the surface density , but only satisfies that condition approximately .
In graphical treatment of the problem this value 100 is the exact datum , whereas in my method we start with an exact value of , and proceed to find the ratio of central to surface density .
With this value of whence I find the foll o series for : \mdash ; .
; ( 16 ) and , by differentiation , the series for is obvious .
This series will be very accurate ) to about .
Thus , or , we have and even the far less convergent series for gives greeing with to the last place of decimals .
When , or , we have , whence , with correct to four , and probably to five , places of , and probably correct to four places of decimals .
This is ] ) ffici e for purpose .
deed , accuracy of this order would be altogether pedantic , it } accumulate .
when I find from this se [ .ies , which lies very to ?
, found bolow .
But the is useless . .
H. ON THE NICAL CONDITIONS OF We cannot , then , rely on this method of procedure beyond the region included etween and , and must now make a new departure .
Since ; therefore , .
( 17 ) Now , let where , after differentiation , is put equal to , a constant .
Then ( 17 ) may be written\mdash ; .
( 18 ) Now , it is clear that Hence , times , we have or or . .
( 19 ) Now , if , for a given value , we know , or , and dyfdx , or , then we can compute from the formula \mdash ; and , by the formula ( 18 ) , .
:\mdash ; be computed .
Afterwards , , \amp ; c. , may be compute by successive applications of ( 19 ) .
This being so , A SWAR .
M OF METEORITES , AND ON OF cosMO G , ( 20 ) In these series may have any value , the series converges adequately .
The convergence may be much improved by an artifice , which , however , I unfortunately did not discoyer until most of the computations were completed .
Let us add and subtract on the right-hand side of ( 20 ) .
Now , If , then , we wlite we have \amp ; c. , , . . .
( 21 ) a more convergent series than that with tlJe A The simplest way of computing the 's appears to be by first the The process for obtaining the numerical solution is then as follows:\mdash ; We have the values of when , that is to say , of when .
From these the successive 's and 's are computed , and the resulting series gives the values of and when is 5 or from this point a new series gives the result when , another series gives the for , and so on .
Later in the calculation several values may be computed from one formula .
* When the computation has been carried out to , we have reached spherical envelope , but that envelope may be by another at any more remote distance from the centre .
Thus , the integration be pursued for values of less than unity , and when the lower limit is zero the envelope is at in finity .
If we write * If the series be far as , several steps be included in one series .
example , the first series , when , may be pushed even as far as without serious error ; for it gives instead of the true value , and , instead of the .
I nob .
ovfever , been satisfied with this egree o racy .
SSOR G. H. DARWIN ON THHi CHANICAL CIONS OF we and But it easily be seen that is a particular solution of the problem ; hence , is a factor by which the particular solution is to be multiplied to obtain the general solution .
The function is given by . . . . .
( 22 ) A table of the values of is given below , showing how the general solution shades off into the particulal solution .
This function , , is also tabulated by RITTER , and I made use of its value , when , to determine the value of , with which the integration is to begin .
I find , however ( see Table I that , when in place of , as given by him .
The last row in Table I. gives the ratio of the central density to , the density at the distance ; this ratio is equal to A SWARM OF , AND ON OF COSMOGONY .
PROFLSSOR G. H. DARWIN THE ICAL CONDITIONS OF It will be noticed that rises from zero to a maximum of about , falls to a minimum of about82 , and then rises to unity .
Since when , we have M. RITTER has for this constant , which he calls It appears from the that density at the centre is times as great as that where ? .
M. RITTER'S solution is intended to that ratio exactly 100 , but this solution shows we to have started with a different value of to obtain that sult .
In the general solution of the diHerential equation the two arbitrary constants may be taken to be the values of when is infinite .
Now , we have taken arily y when is infinite , and the physical conditions of the problem imply that is zero when is infinite .
For if had any positive or negative value different from zero , it would mean that at the centre there was a nucleus of infinitely small dimensions , but of finite positive or negative mass .
Now , is that distance from the centre at which the density is of the central density ; hence , we may regard as the arbitrary constant of the solution .
tever be the sticity of the gas , we may always take as our unit of that distance the centre of the nebula at which the density has fallen to of its value .
Hence , the above table gives the general solution of the problem , subject , however , to the condition that there is no central nucleus .
If we view the nebula from very great stance , appears very small , and thus the solution of becomes .
It is easy to verify that this is a particular solution of the diHerential equation , as is pointed out by RITTER in his ) I fonnd this solution very useful in a preliminary consideration of the problem treated in this The next point which we have to consider is the form which the solution will take , if , instead of taking as the unit of length , we take any other vadue .
The density at any distance and the elasticity are to remain unchanged , but are to be referred to new constants .
Thus , remain unchanged , but are to be referred to , instead of to Now , sillce w remains , since remains unchanged , * I have made of this solution in a paper in the ' Royal jety , it , also been refelled to in a paper by Sir W. THOlJ $ .
Mag vol. 2.3 , .
) .
Sir W. covers much ) same ground as of M. lier papers , but was written hiul ) A SWARM OF METEORITES , AND ON OF COSMOGONY .
From these relations it is clear that and Then , since , when , and since and , it follows that ' when . . . . . .
( 23 ) This relationship has been already used for determining It is obvious also that , when Therefore , ( 24 ) If be density when , we have , when .
( 25 ) If be the pressure when , we have If , therefore , we write , when .
( 26 ) By ( 26 ) we are able to find how the pressure on an envelope of given radius varies with the variation of the telnperatlre of a given mass of gas contained in it .
By means of the formulae ( 23 ) , , ( 26 ) , we are now able to obtain from the number of other ones ; for , after the nges h been ccted in the notation , we may proceed to or diminish the various values of until they are of one size , and we shall thus obtain the solution for at any ) rature whatever .
I shall now proceed to give a table of results when the standard radius , which be conveniently called the , is placed successively infinitely ) centre , where , at ?
so on .
The first of entries gives the various of ( conIputed ) , which the elasticity of the gas depends ; the second line gives ( computed ) , or of DCCCLXXXTX.\mdash ; A. the boundary density to the mean density of all inside of it ; the third line gives ( computed from ( 26 ) ) , by which we trace the variations of pressure at the boundary .
II .
Value of by reference to former Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to Valuc of by refcrencc to The minimum of occurs when very nearly , for , when , I find , 26437 , respectivel When and .
The minimum of occurs when very nearly , for , when , I find , 3462 respectively .
When is a , the density at the centre is 381 times that at the boundary , and , when is minimum , the density at the centre is 129 times that at the boundary .
M. RITTER pressure to be a minimum when this ratio is 258 , instead of 129 .
As this corresponds to , this discrepancy between our solutions is not so large as might be expected from the great discrepancy between these results , I cannot but think that my result is more accurate than his .
The minimum of occurs when ) , and its value is 39723 .
This value makes the surface density exactly one-third of the mean density , for is a minimum when is a ximum , and this occurs when ] ; , when relationship is satisfied , It is i11teresting to note that in this case is very nearly equal to , so that the * Mr. HILL finds that the imum value of to , or 2667 .
The agreement between our results is satisfactory .
A SWARM OF METEORITES , AND ON THEORIES COSNIOGONY .
total internal kinetic energy of agitation of the sphere of gas at minimum temperature limited by the radius is very .
Now , the energy lost in the concentration of a homogeneous sphere from a condition of infinite dispersion is exactly , might , therefore , be suspected that .
is only an approximation to , which may be the rigorous value .
But my numerical calculations were carried out with so much care that I find it almost impossible to believe that there is an error as large as 3 in the third place of decimals , or , indeed , any error at all in the third figure .
Moreoyer , it would be expected that , if this very simple relationship is rigorously correct , it would be possible to prove it rigorously , just as it is rigorously shown above that ; but I am unable to find any analytical relationships by which the minimum value of can be deduced .
If my arithmetical process be correctly carried out , then we ought to find that , when should be equal to or .
Now , I find that , when and , so that the two agree to four places of decimals .
I conclude , therefore , that the true minimum of is 3972 It will be observed that , as increases to infinity , terminates by being equal to .
RITTER has found that it rises above , and oscillates about that value an indefinite number of times with diminishing amplitude , gradually settling down to as becomes infinite .
The values in the preceding table are not , however , carried far enough to exhibit these of .
A consequence of this result is that there are a number of modes of equilibrium of a gas at a given temperature , provided that the temperature lies within certain narrow limits .
This very remarkable conclusion is rendered more intelligible by Mr. HILL'S treatment than by M. RITTER'S .
This point has , however , no bearing on the present stigation .
In any one of the solutions comprised in Table II .
we may complete the table of densities by the formula ( 24 ) , viz. , and I shall later proceed to do this in the one case which has interest for present problem , namely , where the temperature is a minimum , so that is 6264 .
The full numerical results may be more conveniently given hereafter , and it will only be now necessary to indicate how they are to be computed .
When , for example , ; thus , our equidistant values of the density and other functions will proceed by multiples up to and the limit of the isothermal sphere is where When the temperature is a minimum , and we have ; therefore , , and , therefore , if be the value of , when * This is confirmed by Mr. HILL .
His equation is equivalent to , and it appears from his tables that .
Now , , and the reciprocal of is } G. H. -tRWlN ON CJTANICAL COND 1 OF .
Thus , for nple , at the centre , is , and when it is The proportion of which is included in radius is .
Hence , the masses may be computed .
At rt of the gravity is to be found from .
or , expressing in terms of at the snrface , we have , since ( 27 ) The angular velocity of a body moving in a circular orbit at any part of the nebula , and its linear velocity are also easily to be found .
S5 .
in I shall now suppose a sphere of gas of mass minimum temperature is bounded by an here in convective equilibrium , with continuity of temperature density at the sphere of discontinuity of radius .
Let be the mean square of velocity of agitation in the isothermal sphere , and at any other radius Then nroughout the isothermal sphere , but in the layer outside gradually decreases to zero .
be the density and the pressure at radius , and the same things at , radius Then , if the of the two specific heats be that deduced from the kinetic theory of gases , without any allowance for intra-molecular vibrations , we have that ratio equal to Hence , and also Now , the equation for the hydrostatic equilil ) rium of the layer is .
( 2S ) A SWARM OF ORITES , AND OF cosMOGONY .
Let and we have ' , since rigorously , Hence , our equation is .
( 29 ) It is obvious the is the whole mass ( expressed in terms of the mass of the isothermal sphere ) enclosed inside of radius .
The differential equation to be satisfied is .
( 30 ) We have seen in the last section that , and , hence , This equation is not so easy to solve as that in the last section , and I rot succeeded in .
the general law of the coefficients in an expansion .
Nevertheless it is easy to find a series which will do all that is Let be any value of for which we know and , and let Assume Then , if the suffix indicates the value of a symbol when and , we have But PROFESSOR G. H. DARWIN ON THE IvIECHANICAL CONDITIONS OF and , so that , if and are known , is known .
The differential equation ( 30 ) which we have to satisfy is or Now , by \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; . . . . .
( 31 ) And .
( 32 ) By equating the coefFicients in ( 31 ) and ) we are able to determine the .
The law of the series ( 32 ) is obvious , and sufficient of the series ( 31 ) is written down to euable us to find .
We can , however , obtain a good approximation to higher coefficients , because the coefficients in ( 31 ) become relatively unimportant .
A SWARM OF METEORITES , AND ON THEORIES OF MOGONY .
We now begin the solution with Hence , whence I compute With these coefficients I find .
( 33 ) Then , evaluating , and combining the several values by the rules for integration of the calculus of finite differences , I find 1.35 .
( 34 1 .
When , we begin a new series with From these I compute It hat z vanishes when or It follows , therefore , that four equidistant values of lying between and correspond to For the first , where , we have , itnd for the last , where ; and , when , or find ; and , when , or , I find Finding for these four values and co .
them by the rules of integration , I find , when .
( 35 ) We thus see that the mass of the whole system is times the mass of the isothermal nucleus , and radius is times the radius of the nucleus .
The mass of the isothermal nucleus is thus 46 per cent. of the whole .
M. RITTECR , taking the ratio of the specific heats as instead of , says that the proportion is about 40 per cent. .
On Gaseous Sphere in " " } \ldquo ; M. RITTER calls a sphere , with isothermal nucleus and a layer in convective above it , a case of isothermal-adiabatic equilibrium .
Since the height of an atmosphere in convective equilibrium depends only on the temperature at the base , and since the isothermal nucleus in our numerical example is at minimum temperature , the thickness of the adiabatic layer is a minimum , and the isothermal nucleus a maximum .
We are now in a position to collect togetlJer all the numerical esults of the last two sections in a form appropriate for our subsequent vestigation .
It will be convenient to refer all the densities and masses to the mean density and mass of the isothermal nucleus .
ravity may also be referred to gravity at limit of the isothermal nucleus , and velocity to , the mean square of velocity of itation in the nucleus .
A SWARM OF METEOOITES , AND ON -ORIES OF COSMOGONY .
CDbS NDCCCLXXXIX .
ESSOR G. H. DARWIN ON THE MECHANICAL CONDITIONS the of its Distribution in A Sphere of Gas .
We shall consider what would be the distribution of kinetic energy in the if each meteorite ( or molecule ) were to fall from infinity to the neighbourhood where we find , and were to retain that energy afterwards .
This will give the disibution of energy in a swarm of the supposed arrangement of density , if the rate of diffusion of kinetic energy to be infinitely slow , and if there were no loss of energy through imperfect elasticity .
The square of the velocity of a satellite a circular orbit is one half of the square of the velocity acquired by the from to the distance of the satellite from the centre .
If the concentration has as far as radius ?
\ldquo ; and if a meteorite falls from infinity to distance , then , if be its velocity , and the velocity in a circular orbit at distance , in the isothermal sphere , , in the layer .
In these formula , by the definitions of and in the first , and in the second .
From these formulae .
was computed in able III .
The value of or gives may be called the theoretical value of the kinetic energy , because it gives us a measure of the amount of tion of energy by diflusion and loss of by imperfect elasticity , which must belore the whole system can assume the form of an isothermal adiabatic ) .
We will ) go on to consider the total energy lost in condensation .
We have seen that the potential lost the fall of a single meteol.ite is in the ) , and in the outer part .
Now , in the isothermal a spherical element of mass is ' the encl.gy lost by its ] is A SWARM OF , AND ON THEORIES OF COSIIOGONY .
Hence , the whole energy iost in the concentration of the isothermal nucleus is But Hence , the energy lost is .
in an isothermal sphere of minimum temperature , and thus the total lost energy is Again , in the adiabatic layer an element of mass is - and , therefore , the energy lost by its fall from infinity is and the whole loss of energy is the integral of this from to 359 .
When , and when .
Hence Thus , the whole energy lost in the adiabatic layer is Add this t , o the energy found before for the isothermal part , atid the whole lost energy of the system is found to be .
( 36 ) let us eyaluate the total kinetic energy existing in form of tion of molecules .
In the isothermal part it is clearly .
In the adiabatic part it is half the element of mass into the of velocity of agitation integrated through the layer , that is to say , , and , since , we have for the total internal kinetic energy of agitation .
This is rigorously one-half of the energy lost in concentration .
Hence , if a leteor swarm concentrates into this arrangement of density , one half of the is in vaporising and heating parts of the meteorites on impact , and the other half is retained as kinetic energy of agitation .
I find by quadrature that .
Hence , the potential energy lost in concentration , and that part of it which is retained as energy of agitation is .
The whole mass of the system is , and we may , therefore , write these 7548 21767 cmd It is clear then that the average mean square of velocity of tion of the whole system is ) Or , shortly , the average temperature is very nearly of the temperature of isotherlnal nucleus .
It follows from this whole investigation that for any given mass of matter , arranged in an isothermal-adiabatic sphere of given dimensions , the actual velocities of agitation are determinable throughout .
S 8 .
On the " " Sphere of Aciion Whel ] two meteorites pass near to one another , each will be deflected from its straight path by the attraction of the other .
The question arises as to whether the amount of such deflection can be so great that the passage of two meteorites near to one another to be estimated as an encounter in the kinetic theory .
We shall now , therefore , find the deflection of two meteorites , .
with the mean relative velocity , when they pass so close asjust to graze one another .
The mean square of relatiye velocity in the portion is , and this may be taken as the square of the velocity at infinity in the relative hyperbola described between the asymptotes of the hyperbola is the deflection due to this sort of encounter .
Let be the semi-axls and Jltricity of the hyperbola .
Then , if be large , the *M .
gives 741 in place of 755 , but , as already remarked , he uses a different value for the ratio of the specific hcats .
A SWARM OF TEORITES , AND ON THEORIES OF COSMOGONY .
between the asymptotes is ; and , if be the radius of either meteorite , the distance ( when they graze ) is .
Therefore , .
By the law of central orbits Therefore , But , since , we have The unity on the right-hand side is negligible , and , since is the deflection in rees , that deflection is if S be the density of the body of a meteorite , Ss3 , and , therefore , this expression becomes Let us find what must be if the deflection is ; we have We may , for a rough evaluation , take as unity instead of , and suppose to be equal to the distance of Neptune from the Sun ( viz. , cm and , as a very high estimate of the value of S , let us suppose the density of a meteorite is 10 .
Then , since the Sun , grammes , and is about a half of the Sun 's mass , we have Hence , grammes , in round nunnbers .
But the Earth 's mass is grammes , and therefore the meteorites are onetwentieth of the mass of the Earth .
It follows , therefore , that , with such small masses as those with which the present theory deals , the deflection due to gravity is insensible , and we need only estimate actual impacts as encounters .
Hence , the radius of the sphere of action of a meteorite is identic,1 with the diameter of its body .
.30 { ESSOR G. H. DARWIN ON ICAL OF S9 .
On for the bility of to a of rtleteorites .
The question at issue is to determine within what limits the quasi-gas formed by a swarm of colliding meteorites may be treated as a plenum , subject to the laws of hydrodynamics .
The doctrines of the nebular hypothesis depend on the stability of a rotating mass of fluid , and that stability depends on the frequencies of its grayitational oscillations .
Now the works of and others seem to show that instability , at least in a eneous fluid , arises from one of graver modes of oscillation , and the period of the grayest mode does not differ much from the period of a satellite grazing the surface of the of fluid .
Then , in order that hydrodynamical treatment should be applicable for the discussion of such questions of stability , the mean free time between collisions must be small compared with the period of such a satellite .
Another way of stating this is that the mean path of a meteorite shall be but little curved , and that the velocity of a shall be but little changed by gravity in the interval between two collisions .
This must be fulfilled not only at the limits of the but at every point of it .
The condition above stated will be satisfied if the space through which a meteorite falls from rest , at any part of the swarm , in the mean interval between collisions is ] compared with the mean path .
If this criterion is fulfilled , then , in most respects which we are likely to discuss , the swarm will beLave like a gas , we must at present confine the consideration of the matter to this general criterion .
It would be Iaborious to determine exactly the space fallen through from rest , because grayity yaries as the eorite falls , but a sufficiently close approximation may be found by taking gravity constant throughout the fall and equal to its value at the point from which the meteorite starts .
We have already denoted by the value of gravity at any part of the swarm , and have tabulated it in Table ] .
in terms of or Now the mean interval is .
Hence , if be the distance fallen in this time , But and Th erefore , .
( 37 ) The factor has been aboye in Table III .
, and it increases from the centre to the outside .
This criterion may be arded from another point of view , for , if the meteorite be A SWARM OF ThORITLS , AND a circular orbit about the centre of the swarm , is th deflection from the straight path in the mean interval between two collisions .
Then the criterion is that the deflection shall be small compared with the mean path .
We may consider the criterion from .
another point of view , and state that the arc of circular orbit descl.ibed in the mean interyal shall be a small fraction of the whole circumference .
The linear velocity in the orbit is giyen by And the mean interval .
Hence , if be the arc ibed with velocity in time But the whole of circumfrence is Therefore , ' The fictor been tabulated bove , in Table III .
S 10 .
On the It is necessary to make assumptions both as to the mass and the density of the meteorites .
We have a right to assume , I think , that density S is a little less than that of iron , say about 6 , and we put equal to .
Then we , and There is but little information about the size of es ; but , if we the symbol , it will be easy , by merely shifting the point in final results , to obtain results for all sizes .
Thus , if cm .
, grammes ; cm .
, kilogrammes ; if cm .
, , and if cm .
, tonnes .
I shall , therefore , keep in the put it equal to unity in the numerical results .
In the first place , making no as to the density or } of the meteorites , we have Then , by substitution in ( 10 ) and ( 11 ) , we have 1 ( 39 ) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 We will now apply this solution to a case which will put the theory to a severe test .
Suppose that the limit of the sphere of uniformly distributed energy of agitation is nearly as far as the planet Uranus , so that , say .
Then the extreme limit of the swarm is at ; but the orbit of the planet Neptune is at , so that the limit is rther beyond Neptune than Saturn is the Sun .
Now , , I find cm .
per sec. per annum , secondsyears ( 40 ) Introducing these values and putting for , I find per annum kilom .
per sec. 1 ( 41 ) per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 per annum kilom .
per sec. 1 Now we have in Table III .
the logarithms of the several factors , which occur last in these formulae ( 41 ) , at various distances from the centre .
It will suffice for our purpose only to take every other yalue from Table III .
The distances from the centre are expressed in terms of the astronomical unit distance , , the Earth 's mean distance from the Sun .
The mean free pabll is expressed both in { he unit and in kilometres ; and the mean tervals 1 ) etween collisions in days .
The criteria and are , of course , pure numbers .
Table IV .
, as it stands , is applicable to meteorites weighing grammes , by shifting the A SWARM OF METEORITES , AND ON THEORIES OF cosMOGONY .
point one place to the right in the last four rows of entries it becomes kilogrammes , one more it becomes tonnes , and another , thousands of tonnes , and so on .
Iy .
BLEj of Results .
The meteorites weigh ) grammes , and have the density of iron .
The swal.lrl extends to being Earth 's distance from Sun .
The incidence of the several planets in the scale of distance is roughly indicated by the names written above .
The criteria show that , if the meteorites weigh kilogrammes , the collisions are frequent enough , even beyond the orbit of Neptune , to allow the kinetic of to be applicable for such problems as are in contemplation .
For , when the two criteria with decimal point shifted one place to the right ) are and 005 , both small .
But , if the meteorites weigh tonnes , the criteria to be very small , about .
If they weigh , the applicability will ccase somewhat beyond where the asteroids now are .
I conclude , , hen , from this discussion that we are justified in hydrodynamical treatment to a swarm of meteorites from the solar system iginated , even in the earliest stages of the history of the swarm .
This discussion has , of course , no bearing on the fundamental hypothesis that meteorites can glance from one another on impact with a virtually high of elasticity ; nor does it do anything tojustify the assumption a consist approximately of a quasi-isothermal nucleus with a quasi-adiabatic ovel it .
This latter assumption I have been led to by the tionb to hicl ) we llow ) .
H. tWIN ON TH MECITANICAL CONDITIONS OF S 11 .
the Diffusion of the Viscosity .
In order to discuss these questions , it will be well to begin with a simple case of fluid motion .
Consider ] motion , in which there are a number of streams of equal breadth parallel to with velocity , and , interpo]ated tween them , let there be of quiescent fluid ; suppose then that we wish to find motion at any after this initial state .
Let the boundaries of the streams be from to .
Then , if be the velocity , , parallel to at 1 ime , and the kinetic modulus viscosity , the equation of tion is The solution of this being of the form , the complete solution satisfying the initial condition is\mdash ; Now , if we refer time to a period , where , then after a time , which is greater than , the solution is sensibly It is clear that the maximum of ?
occurs when , and the minimmn when and that , they Hence , the difference between the maximum and minimum is 4 .
Therefore , the ratio of the greatest difference of velocities after time to the initial difference of velocities is .
When is 1 , 2 , 3 , this ratio assumes the values resI)ectively .
Thus , after three times interval , the differ YJce of velocities is .
The time may be therefore taken as a convenient measure of viscosity .
In our problem the streanls must be taken of a width compare ble with the linear dimensions of the solar system .
I therefore take , the width of the reams , as equal to , the Earth 's distance the have Now , according to the kinetic theory of gases , the kinetic modulus of viscosity is into the mean frcc multiplied by meall .
Hence , A SWEAR ) OF METEORITLS , ON cosMOC O Hence , we have .
( 42 ) I we apply this formula to the solution which has been already found in Table Iy .
, we obtain the following results:\mdash ; , 16 , , 24 , 3-9 , 108 792 , 193 , , 337 , 27 These applicabL to meteorites ammes in a swarm extending to the meteorites weigh rammes , the values of would be one-tenth of the tabulated values .
If the streams were ten times as broad , the periods would be a hundred times as Now the periods in the above table , eyen if multiplied by a thousand , must be considered as short in the history of a stellar system .
It thus appears that the quasiviscosity must be such that a swarm of meteors must , if revolving , move nearly without relative motion of its parts , at least in bhe early stages of its evolution .
But let us consider the values of at different epochs in the history of the same system .
If be the dius of tlJe isothermal the formulae and ( 10 ) that ries as , whilst ries .
Hence varies versely as Thus , as the Qwarm contracts , the periods increase rapidly .
Thus , later in the history , the viscosity will probably fall offso much that eqnali ation of angular velocity be no longer attained , and we should then the central ting Ipidly than the outside , with a gladual transition one angular velocity to the other .
The ives , besides the , the rate of equalisation of the kinetic energy of tion ; this corresponds in true gas with the conduction of heat .
The conclusion at we thus arrive appears to justify the ) ' the whole of the central of the swarm endued with uniform kinetic energy of tion , and that the nlass of the qnasi-isothermal nucleus is the greatest ) Jble .
to the assumption that the nucleus is coated with a layer in convective , it may be rked t ) the velocity of must decrease when we get to } , outskirts of the swarm , and ) ecti v will satisfy the conditions of the se ttcr than alJy ] ) there c tions will be adduced on in ) Summary .
*blEl Feorio d. The is from a } ical qntnre hich gives the value , and it is ) ntly only accidcntally equal to is the mcnn velocity denoted by MEYER .
SSOR G. H. DAlt l THE CONDITIONS OF S12 .
the Rate of Loss of Imperfect Elasticity , and on the liIeat G In a collision between two meteorites the loss of is pyobably proportional to their relative kinetic ergy before impact .
Therefore , the amount of heat generated by a meteorite .
unit time is to the kinetic ( say h ) and to the frequency of collision .
By ( 10 ) the frequency , or reciprocal of , varies as ; but is to , and varies Hence , the frequency of collisiou varies as , and the amount of heat generated by a single meteorite per unit time ries as ) .
But , if be the quasi-hydrostatic pressure , varies as and , therefore .
the beat generated by a single meteorite varies as Then , to find bhe total heat generated per unit time and , we to multiply this by the number of meteorites per unit volume , that is to say , by , which is equal to Thus blJe amount of heat generated per unit time and volume is proportional to .
With meteorites of uniform size , and with uniform kinetic energy of agitation , this becomes simply the square of the hydrost tic pressure .
The mean temperature of the gases volatilised by collisions must depend on a variety of iderations , but it would seem as if the temperature would follow , or less closely , the variations of heat generated per unit time and volume . .
On the of a of The law of ibution of meteorites found above depends on the frequency of .
But at some distance from the centre collisions must have become so rare that the statistical method is inapplicable .
There must then be a sort of fringe to the swarm , I attempt to represent by supposing that beyond a certain radius ( not the same as the former ) collisions never occur , and each meteorite describes an orbit under gravity .
Now , at any point gravity depends on the mass of all the matter lying inside a sphere hose radius is equal to the distance of point the centre of the swarm .
Hence , the value of gravity depends on the law of density of distribution of meteorites , which is the thing which we are seeking to discover .
We , then , ) every point of a sphere of radius a fountain of meteorites is shot , at all inclinations to bhe vertical , and with velocities grouped about a mean velocity , according to the exponential law appropriate to the case .
As many meteolites are supposed to back on to the surface as leave it , and this cannonade gainst the boundary of the sphere exactly balances the quasigaseous pressure on the inside of the sphere .
Thus , the ideal surface may be annihilated .
Since the falling half of the orbit of a meteorite is the facsimile of the rising half , we need only trace the body from projection to apocentre , and then double A OF METEORITES , AND THEORIES OF )fOGONY .
distribution of density which is deduced on the that all the meteorites are .
Again , since every element of the sphere shoots out silnilar foun , and since collisions are precluded by hypothesis , we need only cider the velocity along the radius vector .
As far as concerns the distribution of density , it is the same as if each element shot up a vertical fountain ; but , of , in ] determining the vertical velocity , we must pay attention to the inclination to the vertical at which the meteorite was shot The mass of the matter inside the sphere , ttraction affords principal part of the force under which the meteorites , is ) , and , for the sake of icity of notation , we shall take as being unit square of velocity .
Now , let be the potential at the point whose radius is , and suppose that a meteorite is shot out from a point on the sphere with a ocity e and at an inclination to the vertical ; { hen , if ?
\ldquo ; be the radius vector and lonoitudea of the meteorite at the tirl ) , the equationls of conservation of moment of molnentum , and of energy are \mdash ; .
If we write , and get Now , we are to regard as the vertical velocity in a fountain squirting up a point on ) sphere .
Then , since , it follows that at the foot of the fountain is qual to .
If , therefore , S be the density at the height and at the foot , the equation of continuity is @ Therefore , But now let us suppose that the meteorites are not ] shot out ab inclination but at all possible inclinations from and It is then clear that this expression must be mnltiplied by , and rated .
Hellae , if 8 now denotes ) integral @ ' SSOR G. H. ] ON THE ICAL CONDITIONS OF where is a constant which it will be unnecessary to determine , and where the limit will be the subject of futu ] consideration .
cting the ration , have S , between limits , It is obvious that , if is greater than , the square root involved in does not vanish for any value of ; and , hence , we must simply take If , on the other hand , ?
is less than this critical value , is value of which l1lakes vanish .
Thus , our formula divides into three , .
:\mdash ; 1st .
greater than ; 2nd .
less than ; 3rd .
less than ; S The physical meaning of this division is as follows : If we take a station near the surface of the sphere , meteorites shot out at all inclinations , eyen tally , reach the height of our station ; and , when they are shot out horizontally , .
If we go , however , to a higher ion , there is a certain inclination which just brings the meteorites at apocentre , where our height ; but those shot out more nearly horizontally to reach us .
Still higher , not even a meteorite shot up lically can reach us , ' nd the density vanishes .
These esult s only espond to a single velocity ; but , if be the mean square of the velocity , the number of meteorites ] ) between and is to Hellae , we have to multiply S by this expression , and integrate to ?
Now , the first term of the first form for S is the same as the second form ; and in bhe third form S is zero ; hence , this term } hen multiplied by the exponential has to be integrated from to .
The second term of the first of @ has to be multiplied by the exponential , rated from to .
Now , for the first put ; * OSKAR , ' Die Kinelische Theory der Gasc , ' 1877 , pp. A SWARM OF IIETEORITES , AND ON THEORIES OF cosLIOGON Y therefore , and the limits of are to Hence , the first term is Again , for the second term put and similarly introduce it into the second term , and we From these expressions we may omit the constant factors ; and , if be the density at height , whilst is the density at the sphere , .
In this unit square of velocity is ; but we have lsewhere written ; hence , if the unit of velocity be given up , we may write in place of , and the result becomes ( 43 ) It is interesting to observe the connection between this law of density and that which would have held if the gaseous law ( due to co]lisions ) had ined .
In that case , since is the potential , we should have had Now , , and , therefore , const . .
, or The first term of our result , then , is exactly ] resulting from the Qeous law , and the second subtractiye term represents the action of the diminished velocity with which the lneteolites move in the higher , when they are liberated from the equalising of continual impacts .
By previous definitio1J , is the excess of the potential at radius above its value at ralius ; hence , .
d. Now , since is required for values of greater than , we may equal to its mean value , between the limits and .
Thus , Hence , If this form for were substituted in ( 43 ) , we should obtain a very complicated differential equation for .
We may , however , find two values of within which the truth must lie .
, if we neglect the action of all the matter lying outside of radius , the second term vanishes , and we have , and the la .
of density is ( 44 ) Secondly , we may supI ) the density to go on diminishing according to the inverse square of the .
We have seen in the preceding solution and tables that this is roughly the law of diminution for a long way side the isothermal nucleus .
According to this .
Hence , in ) second term of we put Hence , , ( 45 ) The substitution of this value in ( 43 ) gives the law of density .
In order to see the kind of results to which these formulae lead , let us suppose , when we reached radius 2 in the adiabatic Iayer , collisions have become so rare as to be rible .
Then the symbols in the ) ulae of this section have numerical yalues ; and , A SWARM METEORITES , AN D ON THEOItIES OF cosMOGONY .
in order to distinguish them , let them be accented , so that , for example , we , \amp ; c. , in ( 43 ) , ( 44 ) , and ( 45 ) , in place of Now Table III .
shows that , -hen ) nearly .
Hence , nearly .
But , at radius in Table 1II .
, } , and this is what we now write or , whilst But very nearly ; herce , , or Thus , Then , substituting for , and noticing that in Table III .
, , when , the first law of density ( 44 ) becomes .
( 46 ) Again , since , and by Table III .
, and nearly .
Thus , according to the second assumption , we have by ( 45 ) , and , since ; and the ] of density is .
( 47 ) The values computed from these alternative formulae ( 46 ) and ( 47 ) be able with those in Table III .
In Table III .
we have the value of computed at distances .
The following short table gives the result extracted from Table III .
for comparison with the values computed from ( 46 ) .
equilib .
, 163 , ) , 033 , First hypoth .
( 4G ) , 163 , 074 , 033 , 015 .
Second hypoth .
( 47 ) , ) 163 , 071 , 029 , 011 .
LXXXI X. lt appears , thel.efore , that the results from the two hypotheses differ but little for some distance outside the region of collisions , and either line may be taken as near enough to the correct result .
We see then that the effect of annulling collisions and allowing each body to describe an orbit is that the density at first falls off more rapidly than if the medium were in convective equilibrium , and that further away the density falls off less ) .
At more note distances density would be found to tend to as the inverse square of this distance .
Thus , the formulae would make the mass of the system infinite .
In other words , the existence of meteorites with parabolic and hyperbolic orbits necessitates an infinite number , if the loss to the system is constantly made good by the supply .
The subject of this section is considered further , from a physical point of view , in the Summary at the end .
S 14 .
On the Kinetic Theo where the Ilfeteorites of all sizes .
In an actual swarm of meteorites all sizes occur , ) , even if this were not the case initially , inequality of size would soon arise through fractures .
Hence , it becomes of interest to examine the kinetic theory on the hypothesis that the colliding bodies are of all sible s , grouped about some mean value according to some law of frequency .
If there be two sets clastic spheres in sucl ] numbers that there are respectively and in umit volume , and if the meall squares of the velocities of the two are and respectively , and if and are the radii of the spheres of the two sets , then it is proved that the number of collisions between them per unit time and volume is We shall now change the notation , and for and write and , and for and write and Then , if S be the density of the spheres , their masses are and The condition for the permanence of condition is that the spheres of all masses shall have the same mean kinetic energy .
Hence , we refer the mass to a mean sphere of radius 9 , and the velocity to a square of velocity , our may be written * The Kinetic of Gases , ' by H. W. A SWARM OF METEORITES , AND ON THEORY S OF COSMOGONY .
But now suppose that there are spheres of possible sizes , and that in unit volume the number whose radius lies between and is Since the integral of this from to is , it follows that is the numbel of spheres of all sizes in unit volume .
be the total mass in unit volume , or the density of distribution , .
If be the mean mass , ; but , ; hence , and If the spheres of radii are those whose radii lie between , and the spheres of radii are those whose radii lie between and Hence , the formula for collisions between the 's and 's is or , if we write , it is .
( 48 ) Tf the spheres are grouped about a mean mass , instead of about a mean radius , according to law of this kind , the subsequent integrals become very troublesome .
Any law of the kind suffices for discussion .
If , however , I had foreseen the tion of S 16 , I should not have taken this law of frequency .
But , and Hence , the number of collisions per unit time and volume between spheres whose radii range between and , and others with radii between and , is The number of collisions of a sphere per unit time is of this , and , since , we have for the collisions of a siJlge sphere the ctor \mdash ; ] nstead of Thell the total number of collisions of all in unit time , or the reciprocal of the mean free time , is the double integral of this from to For the purpose of carrying out the integration , we may veniently , as an algebraic artifice , change from the rectangular axes to the polar coordinates Thus , \ldquo ; Now , if we } ) For tha sformation of the second integral , put we find Hence , the whole integral is and the mean frequency of collision of a single ball per unit time is A OF METEORITES , AND ON ORTES OF COSMOGONY .
The second of these two cannot , I bhink , be evaluated , but its value is easily found by quadratures .
I find , then , The former of the two integrals may be evaluated as follows:\mdash ; Let then , , where is the complete elliptic integral with modulus Hence , We thus have the freqnency collision given by .
Now , ables give with which value we easily find the mean ci , or frequency , nearly .
( 49 ) If be the frequency of collision when the spheres are all of the same size and mass and , and are itated with mean square of velocity , we have , by the ry theory , . . . .
( 50 ) * I owe this to Mr. ORSYTH , the result verifies an luation 1 which I had made .
It follows , therefore , that in our case are more frequent than if the balls were all of the same size in about the ) ortion of 4 to 3 .
In order to find the mean free path , we require to find the mean velocity .
If be the mean of the velocity for any size , the proportion of all the spheres of that size which move with ities l. between and is where But number of spheres of size between and , in unit yolume , is where Hence , the mean velocity is given by Now , , and , or so that , and Therefore , But , and ; therefore , This integral may be evaluated as follows:\mdash ; Let A SWARM OF METLORITES , AND ON TH ORIES OF cosMOG O Now , ' and in the present case where and are the complete elliptic integrals with modulus In LEGENDRE'S Tables , we find , and Then , The mean free path 5 .
and thus .
( 51 ) If the spheres had all been of the same size , we should have had .
( 52 ) Hence , finally from ( 49 ) to ( 52 ) , if there be a nnmber of spherical meteorites , of uniform density , of all sizes with radii grouped about a mean radius according to the law of error , and if be the diameter of the meteorite of mean mass , and be the density of the distribution of meteorites in space , and their mean kinetic energy of agitation , then the mean free path , mean free time , and mean velocity are given by nearly , nearly , } ) nearly .
Also the mean free path is about , and the mean free time about of that which would have held if the meteorites had all been of the same size and had had the same mean kinetic energy * I this to ] PROFESSOR .
II .
] ) ON CIIANICAL CONDITIONS OF S15 .
On Variation of )Collision and Mean Path for the of Each size of ball has its own mean frequency of collision and mean free path , and it is well to trace how the totad meaIlS evaluated in the last section are made up .
We have already seen in ( 48 ) that ( substitutin fur its value in terms of s ) the number of collisions per unit time between ] of sizes to and balls of sizes to is where But the number of balls of size to in it volume is Hence , the mean frequency of collision for a ball of size with all others is Now , Therefore , if we write for the frequency of collision of a ball of size with others , we have Now , the mean frequency for all sizes is given by Hence , .
( 54 ) The involved here cannot in genel.al be detelmined algebraically ; but , if be very small , or very great , we can find an approximate value for it .
If be very small , the integral becomes A SWARM OF METEORITES , AND ON THEORIES OF COSMOGONY .
If be very large , the integral becomes Now , Therefore , the integral becomes and with the known values of and this gives us For intermediate values of recourse must be had to quadratures for nating the integral .
I have therefore determined , by a rough numerical process , sufficient values of the integral to render possible the drawing of a curve for the values of for all values of .
The following table gives the results for the integral , which may be denoted by 1.71 , 4.94 , 12.97 , 28.75 .
If these values be introduced in the mula we obtain 80 , 88 , 1.26 , 1.81 These values are used for .
the curve , entitled " " frequency of collision in below , and they are supplemented by the values found above for , in the case where is either very small or very large .
The frequency becomes infinite when the balls are infinitely small , because of the infinite velocity with which they move , and infinite for infinitely large balls , because of their infinite size .
But it must be remembered that there are infinitely few balls of these two limiting sizes .
We have now to onsider the free path , say , for the several sizes .
If ?
be the mean square of velocity for the size , the mean velocity for that size is , by the ordinary kinetic theory .
From the constancy of mean kinetic energy for all sizes , we have so thaf ) the mean velocity for size is .
But , if be the mean eloci t , and the mean free path , and the mean free time for all sizes together , we have Therefore , the mean velocity for size is But the mean velocity for size is ; hence , When , is very small , we find , and , when is very large , Thus , for small of , the mean free path reaches a constant limit 4 , and for large values it becomes infinitely small .
The intermediate values , sufficient for drawing a curve , are given in the following short table:\mdash ; . . .
A SWARM OF METEORITES , AND ON THEORIES OF COSMOGONY .
These values are set out in the annexed figure in the curve marked " " free path and are supplemented by the values found above for small and large values of The constant limit 4 falls outside the figure .
The horizontal portion of the curve is asymptotic to the -axis .
No immediate use is made of these conclusions , but it was proper to examine point in the theory .
S 16 .
On the Sorting of Meteorites according to size its Resulis .
It is a well-known result of kinetic theory of gases , that if a number of difl'erent gases co-exist , each gas has the same density as though it alone existed , and was subject to the resultant forces of the system ; also the mean kinetic energy of itation of each gas is the same .
From this it follows that the elasticity of each gas is inversely proportional to the mass of its molecule .
Carrying on this conclusion to meteorites , we see that the elasticity of the gas formed by large meteorites is less than that for small ; and , hence , there is a greatel concentration of large nleteorites towards the centre , and there will be a sorting according to size .
The object of this section is to investigate this point .
In SS 14 and 15 , the laws of a kinetic theory investigated when the consisted of molecules of all masses , grouped , according to a law of frequency , about ) a certain mean radius , and molecules of infinite mass were considered to be with , of course , infinite rarity .
Now , if we were to continue to use that law of frequency of masses in the present investigation , we should find , as an lytical result , that the mean mass in the centre of the swarm becomes infinite .
The of very large meteorites in sufFicient numbers to give statistical constancy in a volume which is not a considerable fraction of the volume of the whole swarm is improbable .
We shall , therefore , treat the case best by absolutely excl o very large masses , When such masses occur , they must not be treated stically ; this is a question which I hope to consider in a future paper .
Had I this conclusion when the investigations of the last two sections were carried out , a rent law of frequellcy of mass would have been assumed .
But the results of those sections are amply sufficient to indicate the conclusions which would have been reached with another law of frequency , and , therefore , it does not seem worth while to recompute the results by means of a fresh series of laborious quadratures .
Any law of frequency would suffice for our purpose which excludes masses greater than a certain limit and rises to a maximum for a certain mean mass .
For the present , I do not specif that law } ) recisely , but merely assume that at some radius , which conveniently be taken as that of the isothermal sphere , where , the number of meteorites whose masses ] between and is Sx ; it is also assumed may range from to zero .
The meteorites whose masses range from to may be deemed to constitute .
Suppose that at radius the number of its molecules per unit volume is Sn , its density Sw , its pressure Sp , and let the same symbols , with suffix , denote the same things at radius .
Since all the partial gases are in the permanent state , they have the same mean kinetic energy of agitation , equal to , suppose .
Throughout isothermal sphere , this is constant , and equal , say , to , but varies with the radius in the adiabatic layer over it .
It follows , thel.efore , that the mean square of the velocity of the particular partial gas to is equal to , and the relation between Sp and Sw is @ .
Let be the excess of the gravitation potential of the whole swarm at radius above its value at radius Then , since each partial gas behaves as though it existed by itself , the equation of hydrostatic equilibrium of the partial gas to is The investigation must now divide into two , according as whether we are considering the isothermal sphere or the adiabatic layer .
r1Isothermal Sphere .
Here we have a constant and equal to , and Sp varies as Sw , so that *This is not to be confused with , the mass of the isothermal sphere .
A SWARM OF METEORITES , AND ON THEORIES OF COSNIOGONY .
or Now it is obvious that ; and , therefore , But , by ition of , S ; hence , S. ( 55 ) This is the law of frequency of mass to at radius Now , if be the mean masses at and respectively , and , if we put , we obtain from the same formula , It is also clear that , if be the total density of the swarm at radius .
( 57 ) By the definition of , and in consequence of the supposed spherical arrangement of matter , we have .
If this value were substituted in ( 57 ) , we should obtain a very complicated differential equation to determine , the solution of which is hopelessly difficult .
We may , howeyer , assume without much error that the ) in the integral expressing is the density of meteorites , of which are of the same size , and are agitated with mean kinetic energy .
If this density be written , we clearly have The values of and may be extracted from Table III .
of solutions in S 6 .
Then we have - , suppose , where is rigorously equal to ; but for computing the approximate value is to be employed .
In order to proceed to the evaluation of the mean mass at various distances , we must assume some form for .
assume , then , that .
lt is easy to show that , and M. Hence , the mean mass , and the maximum frequency is for masses equal to Then , by ( 56 ) , we haVe for the mean mass at radius / ) But ( 58 ) It may be remarked that , if be treated as small , we have the first of these integl.als equal to , and the second equal to , and the ratio of the first to the second is .
In order to evaluate , we proceed to introduce the approximate value for ( 1 .
Now , , and ; then , writing for brevity , we have .
( 59 ) A SWARM OF , AND ON THEORIES OF COSMOGONY .
Also , if be small , the approximate result is Before proceeding to give numerical values for the fall of mean mass as we proceed outwards from the centre of the isothermal sphere , we must consider The Adiabatic Layer .
In this case we assume , as before , that the ratio of the two specific heats is , and we therefore have for the relationship between Sp and Sw at radius Hence , But , since , Su)o apply to the radius where , a constnnt , Thus , in the adiabatic layer the equation of hydrostatic eqnilibrium is whence , , ( 60 ) or The investigation now ] lows a line parallel to that taken before .
We have , and , so thab S. This is ) law of frequency of lying between and at radius As Sn can never be negative , we see that there can be no mass greatel than ; and , if be the greatest positive value of the expression , there can be no mass greater than the smaller of or M. Thus , if be the mean mass at radius , ( 61 ) where is the smaller of and M. If we put in 1 ) , obtain , the mean mass at radius It is clear also that , if ) ) the total density of the swarm at radius .
( 62 ) By definition of , and in consequence of the supposed spherical arl.angement of matter , we have If this value were substituted in ( 62 ) , we might obtain a complicated equation for .
It is clear , however , that an adequate approximation may be obtained by assuming that the in the integral expressing is the density of meteorites , all of which are of the same size ?
} , arranged in a layer in convective equilibrium , and with kinetic energy of agitation at the limit equal to If this density be written , and if be the mean square of velocity of agitation at radius , we have , by ( 60 ) , and in consequence of the relationship , and .
Let for brevity ; then , adopting the law of frequency , as before , we have for the mean mass at radius ( 63 ) where is equal to the smaller of and 8 .
A SWARM OF METEORITES , AND ON THEORIES OF cosMOGONY .
The solution now becomes different according as is the smaller .
First , suppose is the smaller .
Then the limits ation a and If we put so that the numerator and denominator of are easily integrable .
If now we write ' ' Then , since , we have .
( 64 ) This expression has a high order of indeterminateness when , but I find when is nearly equal to unity Thus , the mean mass is where , which we know to be correct .
Secondly , suppose that is smaller than M. Then effecting the integrations in the same manner as before , we have , \mdash ; Therefore , or ( 66 ) Tn order to compute from the formulae , ( 59 ) , ( 64 ) , ( 66 ) , it is necessary to make an assumption as to the value of the mass of the meteorites of uniform size whose arrangement of density is supposed to be the same as that of the heterogeneous meteorites .
We have suppose that the law of frequency of masses is known at radius , and that the mean mass is there equal to .
Now , inside of that radius ) masses are more frequent , and outside of it the smaller masses .
I suppose , then , that throughout the isothermal sphere lies half way between or and the maximum , and in the adiabatic layer that it lies half way between or and the minimum mass Thus , inside I take , and outside As we only want to consider the general nature of the sorting process , these assumptions will suffice .
It may also be remarked that a large variation of is required to make any considerable difference in the numerical results .
We now have\mdash ; In the isothermal sphere ( where ) , ; In the adiabatic layer , Tltus , our formulae are:\mdash ; Tn the isothermal sphere , from ( 59 ) , ( 67 ) A SWARM OF METEORITES , AND ON THEORIES OF COSMOGONY .
in the adiabatic layer , when , from ( 64 ) , ( 68 ) when , from ( 66 ) , .
( 69 ) values of and of are in III .
, and from these I compute\mdash ; isothermal .
48 80 1 92 49 38 30 27 24 22 These values ( together with two others in the isothermal part ) are set out in fig. 2 , and show the law of ution of mean mass from centre to outside .
Fig. 2 .
Diagram showing diminuticu of mcan mnss centre to outside .
SSOR G. H. DARWIN ON THL CHANICAL CONDITIONS The uation of mass in the ) , where collisions are supposed to be non-existent , is not very difficult , although it ] volves some troublesome algebra .
I do not give the investigation , merely rking that it leads to almost exactly the kind of of diminution of mean mass as we have in the adiabatic layer .
The first and second JJlS only involved arguments of a genelal character in which analysis unnecessary .
The reader who does not wish to collcern hinlself with details may be supI ) osed to have passed from and 2 to this In order to submit the theory to adequate test , it is necessary to discuss some definite case of the gregation of a swarm of meteorites , and it is obvious that the only system of which we possess any knowledge is our own .
It is accordingly supposed that a number of meteorites have fallell together from a condition of wide dispersion , have ultimately coalesced so as to the Sun and planets as their progeny .
The object of this paper is to consider mechanical condition of the system after the cessation of any considerable supply of meteorites from outside , and before the coalescence of the swarm a star with attendant planets .
the sake of simplicity , the meteorices are considered to be spherical , and are treated , at least in the firsb instance , as being of uniform size .
It is assumed provisionally that the kinetic theory of gases may be applied for the of the distributicn of the meteorites in space .
No account being takerl of the rotation of the system , the meteorites will be arranged in concentric spherical layers of sity of stribution , and the quasi-gas , whose molecules are meteorites , being compressible , the density will be greater towards the centre of the swarm .
The elasticity of a gas depends on the kinetic energy of agitation of its molecules ; and , therefore , in order to determine the law of density in the swarm , we must know the distributi-on of etic energy of lgitation .
It is assumed that , when the comes under our llotice , uniformity of distribution of energy has been attained rrhout a central sphere , which is surrounded by a layer of meteorites with that distribution of kinetic energy which in a gas corresponds to convective In other words , we have .
quasi-isothermal sphere surrounded by what may be called an atmosphere in convective and with continuity of density and velocity of sgitation at the sphete of sepilration .
Since a gas in conyective equilibrium the law connecting pressure and density is that which holds when the gas is contained in a vessel impermeable to heat , such an arrangement of gas has beelI called by M. " " an isothermal-adiabatic sphere and the same term is adopted here as plicable to a swarm of meteorites .
The justifiability of these assumptions will be ccnsidel.ed later .
The first problem which presents itself , then , is the equilibriun ] of isothermal A SWARM OF METEORITES , AND ON ORIE S OF cosMOGONY .
sphere of gas under its own gravitation .
The law of sity is determined in S 4 ; but it will here suffice to remark that , if a given mass be enclosed an envelope of giyen radius , there is a minimum temperature ( or energy of itation ) at which isothermal equilibrium possible .
The minimum energy of gitation is found to be such that the mean square of velocity of the meteorites is almost viz. of the square of the velocity of a satellite , razing the surface of the sphere in a circular orbit .
As indicated bove , it is supposed that in the meteor-swarm the rigid enyelope bounding the isothermal sphere is replaced by a layer of meteorites in convective equilibrium .
The of density in the adiabatic layer is determined in S 5 , and it appears that , when the isothermal sphere has minimum temperature , the mass of the atmosphere is a tively that of the isothermal sphere .
Numerical shows , in , that the isothermal sphere cannot amount in mass to more than 46 per cent. of the mass of the whole isothermal-adiabat-lc sphere , and that the limit of the adiabatic atmosphere is at a distance equal to the radius isothermal sphere .
A\ldquo ; A table of various quantities in such a systen ] , at various distances from the centre , is given in Table III .
, S 6 .
It is ext proved , , that the total energy , existing in the form of energy of agitation in an sphere , is exactly half of the potertial lost in the concentration of the from a condition of infinite dispersion .
This result is bronght about by a colltinual transfer of energy from molar to a molecular form , for a portion of the kinetic energy of a meteorite is ntly being transferred into the form of thermal energy in the volatilised gases generated on .
The thermal energy is then lost by radiation .
It is impossible as yet to sum up all the considerations which go to justify the assumption of the isothermal-adiabatic arrangement ; but it is clear that uniformity of kinetic energy of agitation in the isothermal sphere must be principally broughc bout by a process of diiiusion .
It is , therefore , interesting to consider amount of inequality in the kinetic energy would have to be smoothed away .
The arrangement of density in the isothermal-adiabatic sphere being given , easy to compute what the kinetic energy would be ) ) any part of the swarm , if each meteorite fell from infinity to the neighbourhood where we find it , and there retained all the velocity due to such fall .
The variation of the square of this velocity gives an indication of the amount of inequality of kinetic energy which to be degl'aded by conversion into heat and redistributed by sion in the attainment of uniformity .
This may be called " " the theoretical value of the kinetic energy ' ; it is tabulated in able I , on the line called " " square of velocity of satellite It from zero at the centre of the sphere to a maximum , which is attained nearly half way through the adiabatic layer , and then falls again .
If the radius of the isothermal sphere be unity , then from to 2 the variations of this theoretical value of the kinetic energy * These results had been } , discovered by M. SSOR G. H. DARWIN CHANICAL CONDITIONS OF small Since this " " theoretical value of the kinetic energy\ldquo ; is zero at the centre , there must have been of energy from without , and considerations of the kind show that when a ) lanet consolidates there must be a cooling of the middle strata both outwards and We must now consider the nature of the criteriol ] which whether the eatment of swarm of meteorites is permissible .
The treatment of an ideal plenum of gas leads to the same result the kinetic theory with regard to any phenomenon inyolying purely a mass , when thab mass is a large multiple of the mass of a molecule ; to any phenomenon involving purely a length , when the cube of that length contains a large number of molecules ; to any phenomenon involving purely time , when time is a large multiple of the mean interval between colhsions .
Again , any velocity to be justly deduced principles must be expl.essible as the edge of a cube containing many molecules passed oyer in a time containing many collisions of a single molecule ; nilar statement must hold of any other function of mass , length , and time .
Beyond these nits , wo must go back to the kinetic theory itself , and in it care must be taken that enough molecules are considered once to impart statistical to .
properties .
limits , then , to the hydrodynamical treatment of gases , the like must hold of the parallel treatment of meteol'ites .
The principal question involved in the nebular hypothesis seems to the stability of a rotating mass of gas ; but , unfortunately , this has remained np to now an untouched field of mathematical research .
We can ) judge of probable results from the investigations which have been made concerning the stability of a rotating mass of liquid .
Now , it appears that the instability of a ating mass of liquid first enters the graver modes of oscillation .
In the case of a ting spheroid of revolution the gravest mode oscillation is an elliptic deformation , and its period does not differ much from that of satellite which revolves round the spheroid so as to graze its surface .
Hence , assuming for the moment that a kinetic theory of liquids had been formulated , we should not be justified in applying the hydrodynamical method to this discussion of itability unless the periodic of such a were a multiple of the of the mean time of a molecule of liquid .
rrying , then , this conclusion on to the kinetic of meteorites , it seems probable that hydrodynamical treatment must be lapplicable for the discussion of such a theory as the meteoric-nebular hypothesis , unless a similar relation holds good .
These considerations , although of a very general character , will afford a critel.ioI ) of the applicability of hydrodynamics to the discussion of the conditions of a of meteorites in the kind of problem suggested by the } hypothesis .
the molecules of liquid descl.ibeol.bits about one another , tho .
would bo tbo monn periodic time of one molecule about another .
A SWARM OF METEORITES , AND ON OF cosMOGONY .
In S9 two criteria , suggested by this line of thought , are found .
They measure , roughly speaking , the degree of curvature of the pursued meteorite between two collisions .
These two criteria , denoted and Aj , will fford a measure of the applicability of hydrodynamics in the sense above indicated .
After these preliminary investigations , we have to consider what kind of meeting of two meteorites will amount to an " " encounter\ldquo ; within the meaning of the kinetic theory .
Is it sible , in fact , that two meteorites can considerably bend their paths under the influence of gravitation when they pass near one another ?
This question is answered in S8 , where a formula is found for the deflection of the path of each of a pair of.meteorites , when , .
with their mean relative velocity , they graze past one another without striking .
It appears from the muIa that , unless they have the dimensions of small planets , the mutual gravitational influence is practically insensible .
Hence , nothing short of absolute impact is to be considered an encounter in the kinetic theory ; and what is called the radius of " " the sphel.e of action\ldquo ; is simply the distance veen the centres of a pair when they graze , and is , therefore , the sum of their radii , or , if of uniform size , the diameter of one of them .
The next point to consider is the mass and size which must be attributed to the meteorites .
The few samples which have been found on the earth proye that no great error can be committed if the average density of a meteolite be taken as a little less than that of iron , and I accordingly suppose their density to be six times that of water .
Undoubtedly , in a swarm of meteorites all sizes exist ( a supposition considered hereafter ) ; for , even if originally of one uniform size , they would , by fracture , be rendered diverse .
But in the first consideration of the problem they have been treated as of uniform size , and , as actual average sizes are nearly unknown , results are given in the numerical table for meteorites weighing grammes .
By merely shifting the decimal poirlt one , , or three places to the right the results become applicable to meteorites weighing kilogrammes , tonnes , 3125 tonnes , and so on .
It is known that meteorites are actually of irregular shapes , but certainly no material error can be incurred when we treat them as being spheres .
The object of all these inyestigations is to apply the formulae to a concrete example .
The mass of the system is therefore taken as equal to that of the Sun , and the limit of the swarm at any arbitrary distance from the present Sun 's centre .
The theory is , of course , most severely tested the wider the persio of the swarm ; and , according]y , in the numerical example the outside limit of the Solar swarm is taken at times the Earth 's distance from the Sun , or further beyond the planet Neptune than Saturn is from the Sun .
This assumption makes the limit of the isothermal sphere at distance 16 , about half way between Saturn and Uranus .
The results , applicable to meteorites of grammes , are exhibited in Table Iy .
, S10 .
The velocity of mean square in the isothermal sphere is of the linear elocity SSOR G. H. DARWIN ON THF , NlCAL CONDITIONS of a planet at distance 16 , reyolying about a central body with mass equal to 4 cent. of of the Sun , kilometres per second ; in the adiabatic layer it diminishes down to zero at distance velocity is dependent of the size of the meteorites .
The mean free path betweelJ collisions from 42 , ) kilometres at the centre , to 1,300,000 kilometres at radius 16 , and to infinity at radius .
The mean interyal between collisions ranges from a of a day at the centre , to three days at radius 16 , and to infinity at radius .
The criterion ranges from oo at the distance of oids , to at radius alJd to infinity radius The criterion smallel .
All these quantities are te - times as great for meteorites of kilogrammes , and a dred time , q as great meteorites of tonnes .
From a consideration of the it appears that , with meteorites of kilogrammes , the collisions are sufficiently frequent , even beyond the orbit of Neptune , to allow the kinetic theory to be applicable in the sense explained .
But , if the eorites weigh tonnes , the criteria cease to be very small about distance 24 , : and , if they weigh 3125 tonnes , they cease to be very small at about the orbit of Jupiter .
It may be concluded , then , that , as far as frequency of collision is concerned , the hydrodynamical treatment of a swarm of meteorites isjustifiable .
Although these numerical results are necessarily affected by the conj e values of the mass and density of the meteorites , yet it was impossible to arrive any conclusion whatever as to the validity of the theory without 1lumerical values , and such a discussion as the above was therefore necessary .
If the particular values used are not such as to commend themselves to the judgment of the reader , it is easy to substitute others in the formulae , and so submit the theory to another test .
I now pass on to consider some results of this view of a swarm of and to consider the justifiability of the assumption of an isothermal-adiabatic arrangement of density .
With regard to the uniformity of distribution of kinetic energy in the isothermal sphere , it is important to ask whether or not sufficient time can have elapsed in the history of the system to allow of the equalisation by diffusion .
S11 the rate of usion of the kinetic energy of agitation is considered , and it is shown that , in the case of our numerical example , primitive inequalities of distribution would , in a few thousand years , be sensibly equalised over a distance some ten times as great as our distance from the Sun .
This result , then , goes to show that we arejustified in assuming an isothermal sphere as the centre of the swarm .
As , however , the swarm contracts , the rate of diffusion diminishes as the inverse power of its linear dimensions , whilst the rate of generation of inequalities of distribution of kinetic energy , through the imperfect elasticity of the meteorites , increases .
Hence , in a late stage of the swarm inequalities of kinetic energy would be set up ; thus , there would be a tendency to the production of conyective currents , the whole A SWARM OF METEORITES , AND ON THEORIES cosMOGONY .
swarm would probably settle down to the condition of convective equilibrium throughout .
It may be conjectured , then , that the best hypothesis in the early ' of the swarm is the isothermal adiabatic arrangement , and later an adiabatic sphere .
It has not seemed to me worth while to discuss the latter hypothesis in detail at present .
The investigation of S 11 also gives the coefficient of viscosity of the quasi-gas , and shows that it is so great that the meteor-swarm , if , revolve nearly without relative motion of its parts , other than the motion of agitation .
But , as the viscosity diminishes when the swarm contracts , this would probably not be true in the later stages of its history , and the tral portion would probably rotate more rapidly than the outside .
It forms , however , no part of the scope of this paper to consider the rotation of the system .
In S 12 the rate ofloss of kinetic energy through imperfect elasticity is considered , and it appears that the rate estimated per unit time and volume must directly as the square of the quasi-pressure and inversely as the mean ] ocity of agitation .
Since the kinetic energy lost is taken up in .
solid matter , it follows that the heat generated must follow the same law .
The mean temperature of the gases generated in any part of the swarm depends on a great variety of circumstances , but it seems probable that its variation would be according to some law of the same kind .
Thus , if the spectroscope enables us to form an idea of the temperature in various parts of a nebula , we shall at the same time obtain some idea of the distribution of density .
It has been assumed that the outer portion of the swalm is in convective equilibrium , and therefore there is a definite limit beyond which it cannot extend .
Now , a medium can only be said to be in convective equilibrium when it obeys the laws of gases , and the applicability of those laws depends on the frequency of collisions .
But at the boundary of the adiabatic layer the velocity of agitation vanishes , and collisions become infinitely rare .
These two propositions are mutually destructive of one another , and it is impossible to push the conception of convective to its logical conclusion .
There must , in fact , be some degree of rarity of density , and of collisions , at which the statistical treatment of the medium breaks down .
I have sought to obtain some representation of the state of things by supposing that collisions neyer occur beyond a certain distance from the centre of the swarm .
Then , from every point of the surface of the sphere , which limits the regions of collisions , a fountain of is shot out , in all azimuths and tions to the vertical , and with velocities grouped about a according to the law of : These meteorites ascend to various heights without , and , in falling back on to the limiting sphere , cannonade its surface , so as to counterbalance the hydrcstatic pressure at the limiting sphere .
The distribution of meteorites , thus shot out , is investigated in S].3 , it is found that near the limiting sphere the decrease in density is somewhat more than the decrease corresponding to convective equilibrium .
But at more remote distances the decrease is less ) , and the density ultimately tends to vary inversely as the square of the distance from the centre .
It is clear , then , that , according to bhis hypothesis , the nlass of the system is infinite in sense , the existence of meteorites with nearly parabolic and hyperbolic orbits necessitates an number , if the loss of the system shall be made good by the supply .
* But , if we consider the subject from a physical point of view , this conclusion appeal s. The ejection of molecules with exceptionally high velocities surface of a liquid is called evaporation , and the absorption of others is called condensation .
The general history of a swarm , as stated in S2 , may then be put in different words , for we may say that at first a swarm gains by condensation , that condensation and evaporation balance , and , finally , that evaporation gains the day .
If the hypothesis of convective equilibrium be pushed to its logical conclusion , we reach a definite limit to the swarm , whereas , if collisions be entirely annulled , the density goes on decreasing inverse]y as the square of the distance .
The truth must clearly lie between these two hypotheses .
It is thus certain that even the very small of evaporation shown by the formulae derived from the hypothesis of no collisions must be in excess of the truth ; and it may be that there are enough waifs and strays in space , ejected from other sy stems , to make up for the loss .
Whether or not the compensation is perfect , a swarm of meteorites would pursue its evolution without being sensibly affected by a slow evaporation .
Up to this point the meteorites have been considered as of uniform size , but it is well to examine the more truthful hypothesis , that they are of all sizes , grouped about a mean according to a law of error .
It appears , from the investigation in S14 , that the larger stones move slower , the smaller ones faster ; and the law is that the mean kinetic energy is the same for all sizes .
It is proved that the mean path between collisions is shorter in the proportion of 7 to 11 , and the mean frequency greater in the proportion of 4 to 3 , than if the meteorites were of uniform mass , equal to their mean .
Hence , the previous numerical results for uniform size applicable to non-uniform meteorites of mean mass about a third greater than the uniform mass ; for example , the results for uniform meteorites of tonnes apply to non-uniform ones of mean mass , a little over 4 tonnes .
The means here spoken of refer to all sizes grouped together , but there are a mean free path and a mean frequency appropriate to each si ze .
These are investigated It must also be borne in mind that the very high velocities , which occur occasionally in a medium with perfectly elastic molecules , must happen with great rarity amongst meteorites .
An impact of such violence that it ought to generate a hyperbolic velocity will probably ] cause fracture.\mdash ; Added Nov. 2.3 , 1888 .
] A SWARM , AND cosMOGONY .
in S 15 , and .
various values are illustrated in fig. 1 .
The horizontal scale in that figure gives the ratio of the radius of each size to the radius of the meteorite of mean mass .
The vertical scales are the ratio of the mean free path of any size to that of all sizes together , and the ratio of the mean frequency for any size to that of all sizes together .
The figure shows that collisions become initely frequent for the initely small ones , because of their infinite velocity ; and again infinitely frequent for the infinitely large ones , because of their infinite size .
There is a minimum frequency of collision for a certain size , a little less in radius than the mean , and considerably less in mass than the mean mass .
For finitely small meteorites , the mean path reaches a finite limit , equal to about four times the grand mean path ; but this could not be shown in the figure without a considerable extension of it upwards .
For infinitely large ones , the mean free path becomes infinitely short .
It must be borne in mind that there are infinitely few of the infinitely large and small meteorites .
Variety of size does not , then , so far , materially affect the results .
But a difference arises when we come to consider the different parts of the swarm .
The larger meteorites , moving with smaller velocities , form a quasi-gas of less elasticity than do the smaller ones .
Hence , the larger meteorites are more condensed towards the centre than are the smaller ones , or the large ones have a tendency to sink down , whilst the small ones have a tendency to rise .
Accordingly , the various kinds are to some extent sorted according to size .
In S 16 , an investigation is made of the mean mass of the meteorites various distances from the centre , both inside and outside of the isothermal sphere , and fig. 2 is drawn to illustrate the law of diminution of mean mass .
It is also clear that the loss of the system through evaporation must more heavily on the small meteorites than on the ones .
After the foregoing summary , it will be well to briefly recapitulate the principal conclusions which seem to 1oe legitilnately deducible from the whole investigation ; and , in this recapitulation , qualifications must necessarily be omitted , or stated with great brevity .
When two meteorites are in collision , they are virtually highly ] astic , although ordinary elasticity must be nearly inoperatiye .
A swarm of meteorites is analogous with a gas , and the laws governing may be applied to the discussion of its mechanical properties .
This is true of the swarm from which the Solar system was formed , when it extended beyond the orbit of the planet Neptune .
When the swarm was very widely dispersed , the arrangement of density and of velocity of of the meteorites was that of an isothermal-adiabatic sphere .
Later in its history , when the swarm had contracted , i6 was probably throughout in convective equilibrium .
The actual mean velocity of the meteorites is deterlninable in of given expanded to a given extent .
The total energy of agitation in an isothermal-adiabatic sphere is half the potential lost in the concentration from a condition of nite dispersion .
half of the potential energy lost , which does not reappear as kinetic energy of agitation , is expellded in volatilising solid matter and heating the gases produced on the impact of meteorites .
The heat so generated is gradually lost by radiation .
The amount of heat generated unit time volume varies as the square of the quasi-hydrostatic pressure , and inversely as the mean velocity of agitation .
The temperature of the gases volatilised probably varies by some law of the same nature .
The path of the meteorites is approximately straight , except when abruptly deflected by a collision with another .
This ceases to be true at the outskirts of the swarm , where the collisions have become rare .
The here describe orbits , under gravity , which are approximately elliptic , parabolic , and hyperbolic .
In this fringe to swarm the distribution of density ceases to be that of a gas under gravity , and , as we recede from the centre , the density at first decreases more rapidly , and afterwards less rapidly , than if the medium were a gas .
Throughout all stages of the history of a swarm there is a sort of oration , by which the swarm very slowly loses in mass , but this loss is more or less counterbalanced by condensation .
In the early stages , the gain by condensation outbalances the loss by evaporation ; they then equilibrate ; and , finally , the evaporation may be greater than the condensation .
hout the swarm the meteorites are partially sorted , according to size .
As we recede from the centre , the number of small ones preponderates more and more and , thus , the mass continually diminishes with increasing distance .
The loss to the system by evaporation falls principally on the smaller meteorites .
A. is subject to gaseous viscosity , which is greater the more widely is the swarm .
In consequence of this , a widely extended swarm , if in tion , will revolve like a rigid body , without relative movement of its parts .
Later in its history , the viscosity will , probably , not suffice to secure uniformity of rotation , and the central portion will revolve more rapidly than the outside .
[ The kinetic of meteorites ma be held to present fair approximation to the truth in the earlier stages of the evolution of the system .
But ultimately the majority of the meteors must have been absorbed by the central Sun and its attendant planets , and amongst the meteors which remain free the relative motion of agitation must have been largely diminished .
These free meteorites\mdash ; the dust and refuse of the system\mdash ; probably move in clouds , but with so little remaining motion of agitation that ( except , perhaps , near the perihelion of very eccentric orbits ) it would scarcely be permissible to treat the cloud as in any respects possessing the mechanical properties of a gas .
] The value of this whole investigation will appear very different to different minds .
To it will stand condemned , as altogether too speculative ; others may think that * Added Nov. 2.3 , 1888 .
A SWARM OF TEORITES , AND ON THEORIES OSMOGONY .
it is better to risk error on the chance of winning truth .
me , at least , it pears that the line of thought flows in a true channel ; that it .
may help to giye a meaning to the observations of the spectroscopist ; and that many ' problems , alluded to , may , perhaps , be solved with sufficient completeness to ) light on the evolution of nebulae and of planetary systems .

