]\gt ; INDEX SLIP .
, G. lI.\mdash ; On , Integrals of the Squaxes of : Harmonic Functions .
Phil. Trans. , , vol. 203 , 1904 , pp. I 11-1$7 .
Harmonics\mdash ; lntegrals of Squares of Functio ) Surfaoe of Ellipsoid .
DARWllr , G. H. Phil. , -t , vol. 203 , 1904 , pp. 111 ] V. On the tegrals oof Harmonic ctions .
G. H. , F. , cived 1 , 190 sequel ) in ' ctions , lnely , ' ' On the ) of ] of \ldquo ; vol. 11 ) .
ped .
of , \amp ; c. vol. , pp. 1-1314 .
I shall to ) " " \ldquo ; ' The Pear-shaped \ldquo ; \ldquo ; In " " Hal.lnonics , \ldquo ; the lctions ) ) expressed ) ) formulae are found for the of of the of all the surface harmonics .
Tbese of course required whenever it is to make use method analysis , the tion of them is thel.efore all absolutely essential step ) ) tionH .
The analysis used in the lnination ( of complicated , and is tible of iJnl ) ) } ement .
) ] be taiIled by the lnethods of the esent P , but 1 do not to spend a great deal of time on an merely to prove the ysis .
In Harmonics\ldquo ; the which really subsists of the solid harmonic functions ificed with the object of convenient approximate forms , and I do not thin it would ) possible to such results without this sacriiice But this course the of rendering it difficult to evaluate the } of of ) face 1nonics .
All the harmonic functions to the third inclusive susceptible of rigorous algebraic exJlession indeed the is true of but llot of the functions of ] fourth these CctSe } expressions the integrals should also be enable , ctnd the ) of present papel is to complete the preceding vestigatiou in this It will be well to by a estatenJ e of the notation .
used in " " Harmonics\ldquo ; was convenient the ) essions See Appendix , below , ( 363 .
) involved .
the llotation used in the two later papers seems preferable where the formulae and symlnefi.ical .
In " " nics \ldquo ; the squares of the selni-axes of the ellipsoid were The .
coordinates colllected with ellipsoidal coordinates by The three roots of the cubic were Lastly to between from to In tlJe two pers I ) ' ; for convenience 1 introduced constant ( easily distin guishable from the of evious 1tion ) defined by The of the s of the ellipsoid were then Th diuates became .
The roots of the cubic were .
This is the hich will be used in the paper .
OF ONIC ]UNCTIONS .
If element of , of ) { hc ] ) on to the ) , it ppears f the st the foot of p. 2.57 of\ldquo ; Stability\ldquo ; that / where si ) .
In the previous papers 1 ] ) the ) of harmonic consists or , ( , , , one of the two -functions being ltiPlied by one of the fonr cosine or sin fto[s .
a of typical sefi , the i to be As it will be convenient ) notation , will write these easily intelligible notation .
fn of eveIl f , and therefore may rate of , the limits of and ) nlultiply the result It is clear then that Similar expressionH t to the other of function , we may proceed with this type the others .
This formula shows that the bles are separable , since we substitute for ( without changing the esnlt , the of the integl .
It has been stated above that two of the roots of the cubic equati ) roportional to and .
By the nature of the monic fnnctions it follows that if is l to a certain lctio of ) is ) ) tional to the same function of cos2 .
It follows that if ) , , where is a constant , which for the present we may regard as being unity .
] then , we must have if there is term and a term in ' then there must be a term in , and } .
It that the coefficient of is COFI } of brevi t I call this function , and we may state that one term in the expression is , where indicates the above function the ' integrals .
It follows that \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; \mdash ; ( 1 ) .
Since ' it is clear that Hence if and differ by an odd nulnber , and if they differ by an evell number .
Also Let us write , so that We must now evaluate these functions .
Since , we have differentiation Integrating } ) etween and and multiplying by we have OF IPSOIDAL SURFACE hIONIC FUNCTIONS .
i5 and by ' Multiplying the first of these by the second by } and adding together we have By successive applications of this formula we may reduce any function until it depends ) 2 , , but the result becomes ted few successive reductions .
Now Then ] } ' } But it is well known this of the ) elliptic with moduli and is Hence 2 , It seems unnecessary to the .
) olved in successiye reductions , I results , :\mdash ; .
2 , 4 , These are the only functions of kind -hich needed ) ev tions ( in this paper .
* See for examp U\ldquo ; ' Theory der tischcn 1'p .
When these functions are introduced into ( 1 ) and the terms re-arranged , I find : . . .
7 .
9 ( 1 \mdash ; ) \mdash ; 3 \mdash ; . . .
\mdash ; . . . . .
( 2 ) .
In this result good added llot deducible the table of functions given above , but terln as stated has ) computed .
The laws the succession of tel .
Ins in the six lines clear , but I do not claim that the of the laws is rigorous .
do not ) el.ceive how each series is derived those it , and 1 have no idea how the sel.ies beginning wit } ] vould gu on .
With sufficient patience it would no doubt be possible to deterlnine the general law of the series , but 1 do uot pose to Jnake the atteml ) present , since we have more than enough for the mediate object in view .
This is , of COUl.se , equally applicable to the of the type .
In eHect the required must d the fuIlctions , nd I take the definitions with slight S of " " the } -shaped Fi , \ldquo ; In rdel to use the ) .
analysis it is llecessilly that t of the -fnnctio and the square ) the function should be the same ftlllctions of of .
But in definitions to be used this symmetry does 1lot ) difficulty , which } , howevel , be easily overcome .
If the -function be by the cosine sin ctio by any ctor the will oduce such factors will render of the squares of the sinein the oper mannel It seems to show results found ccord with the .
S " " \ldquo ; this connection I , when written in rJnonics , \ldquo ; is M. lt does nut to fnll details of the analysis in the several OF SOIDAL SURFACE bIONIC FUNCTIONS .
since it is sometimes tedious , it the substitntion in the formula values of , \amp ; c. We will now take the monics successively .
HARMONIC 0F THE ORDER ZERO .
This trmonic is simply unity , so and other 's vanish .
The is . . . . .
( 3 ) .
This obyiously ) , of is the known value .
ITAR 1 Here we all ) unct i oper symmetrical , , I ) .
( 1 ) 1 define this ( 4 ) .
where ] , On ( irlb ) ?
, it Whence I find ince with definition ( 4 ) ( 5 ) .
In ic is defined ; Now take values as , the two definitions into accord .
is the c if ' ; and Hence . . .
( 7 ) , the csult on p. 549 of " " Harlnonics\ldquo ; for the case ( 2 ) 1 define this ( 8 ) , where .
By synnuetry the last result ( 9 ) .
nicH \ldquo ; the functions ( 10 ) .
Nl I ( 11 ) . .
ofnonics \ldquo ; OOC .
( 3 ) rmonic .
I I COh , , , On we ( 13 ) .
In rlnonics \ldquo ; the rlefinitions were the same , and therefore ( 14 ) .
is on of \ldquo ; type HAltNONlCS or the only coefficients and ( si wifl ] , 1 , 2 .
( 1 ) Tl , Cosun These are defined thus , ( 15 ) , where , with upper sign for and lower for ; and Writing , ( 15 ) , , where .
It may be noted that is a syImetrical function in and Squaring we find OFFSSOR .
IT .
ARlVfN O THE OF After reduction find , Now Whence substitution , ( 16 ) .
, the lower for the sectorial harmonic .
Tf these ] ) ) in of as far as three terms of the series , I find , on the ' .
( 17 ) , I ( 18 ) .
In " " Harm \ldquo ; narle t ( 19 ) .
In order to make the two forms of definition must take Thus .
Now on development .
SQUARES OF LLIPSOIDAL S HARMONIC FUNCTIONS .
ence .
Introducing this I find ( 20 ) .
( ' with the result ) .
of " " onics \ldquo ; ) Again in " " Harmonics\ldquo ; ' ( 21 ) .
In order to make the two definitions we must take ; or , .if So .
Introducing this in ( 18 ) we have .
1 .
, ( 22 ) , eeing)with the result on p. 548 of " " monics \ldquo ; with , type LEC .
( 2 ) Cosine These are defined thus , \mdash ; .
( 23 ) where Squaring we find On substituting in the formula , find , on putting COS ( 24 ) .
VOL. CCIII.\mdash ; A. In " " Harmonics\ldquo ; the definitions were ( 25 ) .
In order to make the two definitions agree we must take ; so that .
On multiplying ( 24 ) by this factor , we have .
3 , ( 26 ) with the result on p. 549 of\ldquo ; Harnlonics\ldquo ; with , type EOC .
( 3 ) The Tesseral Sine rmonic .
This is defined thus , \mdash ; sin .
( 27 ) where 9 , we find whence , on putting ( 28 ) .
In Harmonics\ldquo ; the debnitions were ( 29 ) .
Therefore , to make the two definitions , we must take SOIDAL S HARMONIC FUNCTION I 2.3 Therefore - .
, ying ( is factor we have , ( 30 ) agreeing with the result on 1 ) .
54 of Harlnonics\ldquo ; with , type EOS .
( 5 ) The rmonic .
This is defined thus : .
( 31 ) .
If in the last integral we had written for 40 , and for , and for would have become and would have become erefore the result ( 28 ) what is lleeded by lnerely interchanging and Therefore . . . . .
( 32 ) .
For the purpose of comparison I must put , ( 31 ) and .
In " " Harmonics\ldquo ; the definition was @ .
In order to make the two definitions agree we must take Thus ; introducing this in ( 33 ) we have .
( 35 ) yreeing with the result on p. 548 of ' ' Harmonics\ldquo ; with , type HARMONICS THIRD In these the only coefficients are , and ( 2 ) becomes , 1 , 2 , 3 ) .
( 1 ) The Zonal Second Tesseral Cosine Harmonics .
These are defined thus : ( 36 ) where , with the upper sign for and the lower for .
Writing ( 37 ) where Squaring we find After some rather tedious reductions I find Now writing OF ELLIPSOIDAL SURFACE HARMONIC FUNCTIONS .
On in HHio n be unity with the dehnition adopted , 1 hnd ( 38 ) .
same with the of changed .
If these expressions be developed in powers of , and if the factor be reintroduced , I find . . . .
In " " \ldquo ; I defined , .
In order that our previous definition may agree with this we must have . .
Now whence , and this value of .
satisfies the second equation .
With regard to there is a mistake in the table ( the only I have detected therein ) on p. 556 of\ldquo ; Harmonics for the coefficient of the second term should rot be but .
The mistake obviously arose from my using the for instead of that for as given on p. 490 .
With the corrected coefficient the definition is ( 40 ) .
In order that the previous definition may agree with this we must have But , and then ce .
This value of will be found to give the correct value for Then , and .
this into the value of , we find ( 41 ) agreeing with the result on p. 549 of\ldquo ; Hal.monics\ldquo ; for , type in " " Harmonics\ldquo ; I defin ed ( 42 ) .
To make the former definition agree with this we must take , .
In the present case omitting the term in we find , with this value of , and that the second equation is satisfied .
Again I defined Hence to secure agreement we must take , Now , and therefore .
second equation is satisfied .
SQUARES OF ELLIPSOII)AL HAltMONIC FUNCTIONS .
We have then .
Introducing this into the value of , we find 3 .
4 .
5 ( 44 ) , eeing with the result in p. 548 of\ldquo ; Hal.nonics\ldquo ; for , type OEC .
( 2 ) First rmonic ctnd These are defined thus : ( 45 ) , where with for lowel sig for , and Writing ( 46 ) , where It is clear that has the s form as when in the latter we interchange with , nd with .
The of the variables of integration clearly makes no difference in the result , therefore we need only interchange and , and replace by In the present instance This shows that is the same function tlJat t was of but that is analogous with , and with .
Thus we once down the results by and throughout .
Let Then putting , we have by with ( 38 ) .
I.33 same with the of changed .
SSOR G. H. DARWIN ON THE GRALS OF If these ) essions be in of I find , on reintroducing ) ' the ft .
, . . . . .
In " " monics " " I defined .
But we have it above by . . .
Now Whence value also satisfies the expression I defined . . .
( 49 ) .
But we defined it above by . .
the value we have ; whence .
SQUARES OF LLIPSOIDAL SURFACE HARMONIC FUNCTIONS .
Therefore , and I find .
( 50 ) , with the result on p. 549 of\ldquo ; Harmonics\ldquo ; with , type OOC .
In " " Harmonics\ldquo ; I I 5 ( 51 ) .
But .
Therefore 15 , .
15 Now Therefore \mdash ; , and This ives the corl.ect value to .
Again . . . . .
But .
Therefore and .
liminate , these equations correct value for Then .
Therefore .
Hence we find 360 ( 53 ) , agreeing with the result on 1 ) .
548 of " " rmonics \ldquo ; , type OOC .
VOL. CCIII .
( s ) First Sine Harmonic and Sectorial Sine Harmonic .
are defined thus : , ( 54 ) where , with the upper sign for , the lower for , and riting , sin , ( 55 ) where Squaring we find , substitution in the formula rmonics of the third order I find If we write On substitution 1 find , on putting ( 56 ) the with of the square root reversed .
SQUARES OF SURFACL IONIC F these in powers of oducing the and reverting to the notation of ' HarInonics , \ldquo ; I find .
In ( Harmonics\ldquo ; defined COfI ( 57 ) .
To make our former definition agree this we must take ; .
Hence It will be found that , aboye form .
Again I defined sin ) , . . .
( 58 ) .
To make our former definition agree with this we must take .
It will be found that .
Whence , and has the correct form .
Therefore .
Whence .
6 ( 59 ) , agreeing with the result on p. 548 of " " Harmonics\ldquo ; with , type OR .
H. RlVIN ON THE ' In " " Harmonics\ldquo ; I deflned , In order that may with our former definition we must take .
Whence .
5 .
It will ) found so the correct form .
Again I defined . . . .
( 61 ) .
In to former definition agree with this we must take Therefol e It will ) , so that has the correct form .
Then , and .
Whence .
( 62 ) , with the esnlt on 548 of ' Harmonics\ldquo ; with , type OOS .
( 5 ) Second This is defined thus:\mdash ; where OF qoIDAI , HARMONIC FUNCTIONS Squaring we have Therefore Reducing this expression and ] ) ] , we have ( 64 ) .
In " " ITarmonics\ldquo ; defined si 1 .
( 65 ) .
To make our former ition agree with this we must take Again I defined .
To make the former definition agree with this we must take Therefore and ' .
Hence in the notation of\ldquo ; Harmonics\ldquo ; 3 .
4 .
5 , ( 67 ) agreeUng with the result on p. 548 of " " Harmonics\ldquo ; with , type OES . .
H. ON OF It llay be convenient , as furnishing a kind of index to the foregoing investigation , to state that the 1 integrals for the harmonics of orders , 1 , 2 , 3 are iven in equations 3 .
5 , 9 , , 16 , 24 , 28 , 32 , 38 , 47 , 56 , 64 , corresponding to the definitions contained in 4 , 8 , 12 , 15 , 23 , , 31 , 36 , 45 , 54 , 63 .
The defJnitions of the harmonic ) as given in my paper on Harmonic repeated , 10 , ] , ] , 25 , 29 , 34 , 39 , 40 , 42 , 43 , 48 , 49 , ) , 52 , 57 , 58 , 60 , 1 , 65 , C6 .
espon d to these latter definitions the approximate are given in 7 , 11 , 14 , 20 , 22 , 26 , 30 , 35 , 41 , 44 , 50 , 53 , , 62 , 67 ; and the confr the correctness of the general approximate formulae for the integrals given inS of the ) .
on H.armonic A mistake in that ) was detected in the value of the cosine-function for the third / the corrected alue is given in ( 40 ) .
lt must be obvious that the method exhibited here may be applied to higher harrnonics ) tever d of accuracy is desired ; but it is also clear that the bour of the eases very much as they rise in order .
It is the ) results of the previous paper will suffice for most practical applications .
NDIX .
the Symmetry of the Cosine Sine-functions with the -functions .
In my previous papers I fniled to notice that the symmetry between the -functions and the cosine and ctions is not destroyed , ) is only masked , in the approximate expressions for the harmonic functions . .
example , ( 39 ) shows us that , therefore , in consequencc of the symmetry which subsists , we ought to find Now , .
Whence .
This only diffel .
S by constant factor from the expression ( 40 ) .
It wonld ] ) then to only o1le type of function , or to express all the cosine and of the } ) one of them .
This would be found to be equivalent to exprcssing the ions in terms of powers of .
For the purposes of practical application I ( think this wonld be so ) venientclS the nse of cosines and sines of multiples of ) , and the SQUARES OF ELLIPSOIDAL SURFACE HARMONIC FUNCTIONS .
advantage of using only one type of function would not compensate for the loss of convenience in the resuIt .
Accordingly I do 11ot think it worth while to undertake the laborious task of revising all the analysis of ' ' Harmonics\ldquo ; from this point of view .
I may mention , however , I have far in this direction to feel pretty confident that , if this new form of developing the and sine-functio1ls were dopted , the coincidence , mentioned in the footnote on p. 547 of " " Harmonics as to the form of the integrals of the squares of surfacs harmonics would become explicable .
POSTSCRIPT .
[ December 24th , 1903 .
] Mr. HOBSON has shown me how ls may be evaluated by a simpler method of analysis , without the intel.vention of elliptic integl.als .
As an example of the method he suggests I take the eyaluated above .
The solid ellipsoidal harmonics are given , except as regards a factor , in S 3 of ' ' The Pear-Shaped Figure In ( 19 ) of that paper we find where is the factor to be evaluated so as to gree with the definitions The ellipsoid over which we desire to integrate is defined by , and extremity of axis is defined by Hence at this poin But at the extremity of the axis Therefore ( ( coscc-7 .
Therefore , and Let us assume ' Then when , is on the ellipsoid we we regard as coordinates of a point on a sphere of unit radius , or as direction cosines , if it is more convenient to do so .
On substitutin for their values in terms of we find On performing the same operation to the points on the boundary of an element of surface of the ellipsoid , we find where is an element of the surface of the sphere of unit radius , or an element of solid angle .
Since on the surface of the ellipsoid ) , it follows that Hence It is easy to prove that Therefore On substituting its value , , and effectin reductions we al.rive at result given in ( 16 ) above .
SQUARES OF ELLIPSOIDAL HARMONIC FUNCTIONS .
It is obvious that this process is considerably simpler and more elegant from the point of view of theory , but to carry these operations through for all integrals given aboye would entail a good deal of lgebra .
I think indeed that the work be very much less than what I have already done .
Mr. HOBSON has further remarked that all the integrations may be avoided by the followin theorem:\mdash ; If be a solid spherical harmonic function of of degree .
Considering , however , how nple are the involved in his first , it may be doubted whether would save ouble .
\mdash ; VOL. CCIIL \mdash ; A.

