]\gt ; 1893 .
] On Operators in ' sical illathematics .
These groups are\mdash ; I. Methane .
IT .
The three methyl compounds .
III .
Ethane and its derivatives .
IV .
Propane and its derivatives .
If the members of a group have the same ratio of the specific heats , we know , a well-known equation in the kinetic theory of gases , that the ratio of the internal energy absorbed by the molecuIe to the total energy absorbed , per degree rise of temperature , is the same for all .
Hence we have the result that , with the single exception of marsh gas , the compounds with similar formulae have the same energy-absorbing power , a result which supplies a link of a kind much needed to connect the graphic formula of a gas with dynamical properties of its molecules .
From the conclusion we have reached , it f'ollows with a high degree of probability that the atoms which can be interchanged without effect on the ratio of the specific beats have themselves the same energy-absorbing power , their mass and other special peculiarities being of no consequence .
there , the anomalotls behaviour of methane c.onfirms what was clear from previotls determinations , namely , that the number of atoms in the molecule is not in itself sufficient to fix the distribution of energy , and suggests that perhaps the tion is the sole determining cause .
If this is so , it follows that ethane and propane have the same configuration as their monohalogen derivatives , but ) methane diifers from the methyl compounds , a conclusion that in no way conflicts with the symmetry of the graphic mulae of methane and its derivatives , for this is a of reactions , not of form . .
" " On tors i Physical tics .
II By OLIVER HEAVISIDE , F.R.S. Received Jutlc 8 , } .
gebraical Eof the Forms of the Fundamental Bessel Function in Ascendinf and Descending Scries by means of the Generalized Exponential .
27 .
As promised in S 22 , Part I Proc , I will now first show how the ] for the Fouriel-Bessel function in rising and descending powers of the variable may be braically harmonised , without analytical operations .
algebraical sion is to be effected by means of the generalized ) theorem , S 20 .
It lvas , indeed , used in S 22 to generalize the ascending form of the function in qnestion ; but that use was analytical .
At present it is to be algebraical only .
Thus , Jet .
O. Heaviside .
[ June 15 , , ( 1 ) , ( 2 ) .
( 3 ) Here A is the usual form the Fourier-Bessel funcbion ( or , rathel the function instead of the oscillating function , whose theory is less easy ) , or the soIution in .
powers of of the differential equation , ( 4 ) as in ( 71 ) , , I. Also , is a particular case , , ( 78 ) , Part I , of the generalization of the same series , ( 77 ) , Part I , using the odd powers of , and going both ways , in order to complete series .
And is an equivalent of the same function in a descending SGl.ies , ( 31 ) , Part I , obtained analytically , before the subject of generalized differentiation was introduced .
The analytical transformation from A to was considered in S 14 .
The preseut question is , what relation does bear to A and algebraically P It cannot be algebraically identical with eithel .
of hem alone , on account of the radical in C. We may , however , eliminate the radical by employing the particular case of the generalized exponential that will introduce the radical anew .
Thus , ( 63 ) , ( 64 ) , Part I , .
( 5 ) If we use this in ( 3 ) , and out the multiplications , we obtain a series in integral powel .
S of , positive and negative ; thus , .
( 6 ) 28 .
Now involves all the odd powels of , whilst A involves only the even positive powers .
But the terms involving even ative 1893 .
] On Operators in Physical Mathematics .
powel .
S in A zero , if we follow the law of the coefficients .
So is also complete , and must be some tion of series and B. fact , if we assumo that is a solution of the characteristic ( 4 ) , and insert it therein , to find the law of the coefficients the manner , we find that the even zero , whilst the even 's are connected in one , and the odd 's and evell b independently connected in another way .
This makes , ( 8 ) lvhere and are independent multipliers .
from common experience with this rule-of-thumb method of constructing soIutious of diffel'ential equabions , we might stily conclude that itnd epresented the two independent solutious of the characteristic .
Here , however , we know ( analytically ) they are not independent , but are equivalent .
Therefore , ( 9 ) where the sum of and is ullity .
It only remains to find the valuc of Ihis is easily obtainable , because the separate scries in ( 6 ) are rapidly convergcnt .
But we need only the first series , to find the coefficient of ] , the first of ( 6 ) gives .
( 10 ) We see , therefore , that the series is algebraically with half the sum of the series A and B. To further verify , we see that coefficieltt of in ( G ) should times that of This requires or which is also close .
Similarly , from the .
we cquire .
O. Heaviside .
[ June 15 , The numerical tests in this example are perfectly satisfactory ; and if the numerical meaning of a gent s could be always as easily fixed , it would considerably cilitabe the investigation of tho subjecb .
Condensed eneralized Notaiion .
of the Descendinj Serres for the Bessel Function through the Generalized Binomial Theorem .
29 .
Since the series A and are particular cases of the general formula ( 77 ) , Part I , or of ( 11 ) by taking and respectively , it may be desirable to find the general formula of which the series is a particular case .
Notice , in passing , the notation employed in ( 11 ) .
It is certainly easier to see the meaning of a series by inspecting the written-out formula containing several terms , when one is not familiar with the kind of series concerned .
As soon , however , as one gets used to the kind of formula , the writing out of several terms becomes first needless , and then tiresome .
The short form ( 11 ) is then sufficient .
One ternl only is written , with the summation sign before it .
The other terms are got by changing wibh unit always , and both ways .
The value of is arbitrary , though of course it should have the same value in every term so far as the fractional part is concerned , so that , in ( 11 ) , may be changed to any other number without affecting its truth .
Similarly , the exponential formula may be written ( 12 ) with bitrary and unit step .
Now , to find the generalized formula wanted , we have , by ( 25 ) , Port .
( 13 ) xpand this according to the particular form of the binomial eorem got by taking in ( 84 ) , Par6 I , arbitrary .
Or iting th general thus : \mdash ; .
' which is compact and intelIigible , according to the above , and write 2 in place of .
This makes 1893 .
] On Operators in Physical Mathematics .
Effect the integration , and we obtain immediately ; and tlrrefore , by ( 18 ) , Here we see the great convenience in of the condensed llotation .
At the samo time , it is desirable to expand sometimes and see what the developed formula looks like .
We then take the written term as a central basis , making it a factor of all the rest .
Thus , .
( 18 ) 30 .
in this , and we have ( 19 ) is the same as ( 20 ) , Part I , noting that there is here .
But of course the expollential factor is now of no service , dinary series , equation ( 1 ) above , being the practical formula when x is small .
Take in ( 18 ) , and we obtain which is the formula , equation ( 3 ) above , the practical formula when x is than is suitable for rapid calculation by A. Observe that these are the cases , for the whole of the second line in ( 18 ) out to make ( 19 ) , and the whole of first line , excepting first term , goes out to make ( 20 ) .
the other , it happens that extreme cases of a generalized formula are numc ] ically uninterpretable .
to the form , we use exponcntial expansion in the form , but with llegatived , thus , .
( 21 ) Mr. O. Heaviside .
[ June 15 , Employing this in ( 18 ) , we can reduce the to one integral powers only .
The coefficient of is made to be ( 22 ) That this reduces correctly to a convergent series summing up to when , may be anticipated and verified .
, that when we unity is sufficiently evident .
In these conclusions we merely corroborate the preceding .
But I have not been able to reduce ( 22 ) to a simple formula showing plainly in what ratio the formulae A and involved when has any other than and integral value , same .
The Extreme Forms of the Theorem .
Obscnrit 81 .
There are some peculiarities about the extreme forms of the binomial theorem when the exponent is negative unity ( or a negative integer ) which deserve to be noticed , because they are concerned in failures , or apparent failures , which occur in derived formulae .
These peculiarities are co1mected with the vanishing of the inverse factorial for any negative value of the argument .
Thus , in .
We obtain Now , on the left side we have the vanishing factor .
So , the right side , the quantity in the brackets shonld vanish .
This asserts that .
, where on the left side we have result of dividing ] by , and , the right , the result of by , or by These series are the extreme forms of the expallsion of by the ordinary binomial theorem , and they are asserted to be ally equivalent , although the numerical equivaIence , hich is S0llletimes recognisable , is often scarcely imaginable .
But observe that if we choose as well , we have a factor on the side also of ( 24 ) .
It is apparently the same as 1893 .
] On in lfathematics .
the other , and coul\amp ; be removed from both sides if it were finite .
It musb not , however , be removed from ( 24 ) .
What is asserted is that , where the first on the is , and also the on the left .
, if we put first in ( 2.3 ) , . . . .
, and then put .
But if we multiply ) by making , ( 27 ) we see that the descending series vanishes when n is any ative .
That is , it is asserted that , ( 28 ) 1iless is negatively .
But when it is a negative there are additional terms , though always in indetel.minate form ; for instance , when and is finite .
It would appear , however , that the value is zero , because there is eve eason to think correct ( as a particular form ) in the limit .
On the other hand , if we multiply ( 26 ) by , and so make it cancel the , \amp ; c. , in the denominators , get , is , ( 29 ) which is quite inadmissible , since ight lnember is the sum of two series iously found to be equivalent to one another , and to the left member .
The right member is therefore twicc as the left .
of the Binomial tcgral Ncgave Index .
32 .
A consideration of the above ities sgests the following way of avoiding them .
We should } that the zeros and , when we take ?
and indepcndent , have any ratio we please .
Thus , { irst put in ( 23 ) , 112 Mr. C. [ June 15 , .
( 30 ) This being genera ] , let and be both infinitely small , but without any connexion .
We know that the rate of increase of the inverse factorial with is 1 when is -1 .
It follows that .
( 31 ) These , used in ( 30 ) , make it become .
( 32 ) Ultimately , therefore , we obtain in a clea manner This seems to be the propsr limiting form of the binomial theorem when the index is negative unity .
It asserts that the two extreme equivalent forms may be combined in any atio we please , since may have any .
If , we have the ascending series only .
If , then the descending only .
If , we obhain half their sum .
The expansion is indeterminate , but the of minateness appears to be merely conditioned by the size of the ratio We also notice that the suppositions that is infinitely small and is nite , so that , and ' used in ( 30 ) , lead us to ; ( 34 ) that is , the difference of the two extreme equivalent sel.ies divided by , which is , of COUl'se , indeterminate .
1893 .
] Operators in sical natics .
Consideration of a more general Operator , .
Suggested Derived 33 .
Some since , after noticing first the analytical and then the Jjumerical equivalence of the diffcrent fortuulae for tho -Bessel function arising iately from the operator by the use of the two extre1ne forms of the binomial eorcm ( only bhen known to me ) , I avoured to extend the resnlts by ting the includes the former , and comparing the orlns .
) , calling the ssrics in ascending powers of , and the descenling , so and tegrating ( with for operand , as usual when no itten ) , we obtain , ( 37 ) ; and the suggestion is that these are equivalent .
If this equivalence analytical , and we substitute for and a second time , we obtain .
( 39 ) ; ( 40 ) and obvious repetitions of same procoss lead us to .
( 41 ) , ( 42 ) hich are clearly the cases and of the general expresbion ; ( 43 ) Mr. O. Heaviside .
[ June 1 is not a negative , when we know that closer examination is required .
ailure of Numerical alence in certain Cases .
34 .
Now , although the equations following ( 35 ) , ( 36 ) ( excepting ( 43 ) ) are deducible from them by the process used immediately and without trouble , there is considerable difficulty in finding out their meaning .
Considering ( 37 ) and ( 38 ) , I knew that in the equivalence was satisfactory all round , though not very understandable .
When is , or integra ] , it is satisfactory , for then we have merely a perversion terms in passing from to But when I tried the case , and subjccted it to numerical ation , with the expectation of finding numerical equivalcnce to the extent permitted by the initial ergence of the divergent se , I found a glaring discrepancy between and on taking , we produce , ( 44 ) , ( 4 which show no sort of numerical equivalence ever .
] arly , gives , ( 46 ) , ( 47 ) which ] do not show any numerical equivalence .
I was therefore led to think that the equivalence in the case of the Fourier-Bessel function was due to some peculiarity of that function , and it is a fact that the function is the meeting-place of many remarkabilities .
The lnater was therefore put on one side for the time .
But , more recently , independent evidence in other directions showed me that there was no particula ] .
reason to expect such complete failure .
And , in fact , on returniog to thediscl.epant calculations relating to , I found an importaYlt nlmerical .
When corrected , the results for and reed as fail .
lie as be expected .
1893 .
] On Operators in matics .
Probable of Numerical Equivalence by Initial within a Range for , viz. , to 35 .
Thus , in ( 37 ) , ( 38 ) produoes .
' .
( 49 ) Here take .
Then was the test case which failed , er } or th lality of two terms , then , a little of another two consecubive , whicl1 caused a now make Their equivalence reqoires that is about right .
When , we have . . . . . . .
And when } ; Mr. O. Heaviside .
[ June 15 , ving quiring Of course with such a small value of , we cannot expect more than a very agreement , becanse the convergence of the series is confined to the first and second terms , and we may expect an error of magnitnde of the ratio of the second to the first term .
36 .
Now take .
We have ; and in case of we have ( 54 ) ; ( 55 ) hich make ; and therefore Now this shows a large error , for the value is about .
This xcess in is , however , made a deficit by not counting the smallest term in series ( the third term ) .
omitting it , we make and , with , we have making ) , This makes which ) is very good .
1893 .
] On in Ph 37 .
Now passing to the case of a bigger , wo may remark that this differs from the good case by an diffel.entiation , so we expect good results .
We ( 56 ) Taking firsf , giving , ( 59 ) we find ) counting last convergent , is , the smallest term in the scries .
Its makes eciably too big , Nexb take .
Then ( 60 ) ; ( 61 ) not counting the smallesb term .
Lastly , with , we have ; giving agoin neglecting the smallest term in it is of little moment in this example .
The tendency for to be too bjoo .
when the smallest , term is fully should be noted .
38 .
A further increase of to gives good ] , and likewisc .
Thus , for we have ; Mr. O. Heaviside .
[ June 15 , giving in the case of ; Thus we have practically gone over the ground from to good results , so far as the limited examples are concer1led , and can be , so far , cely a doubt of the existence of numerical equience , iu the same sense as before with respect to the ascending and -Bessel fnnction .
Itl.emains to examine between and -1 .
This is important on account of the complete failure in the latter case of the numerical when timated in the above manner .
From the ready shown indeterenCSS of the binomial nsion when , we have the .
of a partial explanation , because we should Tive at the , where .
But there remains the facb indicated that the extreme forms of the binomial expansion are , so that we should expect and to be equivalent .
Since , wever , the numerical equivalence of the different forms of becomes very when is or is near , so we should not be surpl.ised to find the Ullsatisfactoriness becomes emphasized in the case of and .
Such is , in fact , the of of Derived Series reckoned by Initiaf ergence , at first slight , later complete , approachei ) egative Integer .
39 .
Take in ( 37 ) , ( 38 ) .
Then ( 66 ) When , we find that as we do not , or do , count the smallest term in .
Tl at is , the first gives far too great a result , whilst the other , though not bad , is still too great .
That the ives too snlall i , when the smallest ternl is fully .
A of tcl .
Ill is needed , to come to On in Ph ical ematics .
WheIl x we deduce that the first case ) .
without , the second with ] , the smallest term in .
Both resuIts are too great , the error is less than the last counted .
But this when we pass to hetl we conclude that the former case without , and latter with ] , the term in the result is too big , and error rule jUbt nlentioned fttils .
For if we add on the term a second time , we , which is still too 40 .
Since the case is bad , we may expect to be .
We have .
( 69 ) Here take , theu that Bub it caltnot lie these limits , being only a ]iftleo1 er unity .
add on to .
next , the third in the serits .
ill is still too great , and , of coursc , the rulc is ong , } pocted just no Whilst there does not to be any departure from cquivalencc of and in the sense used } ) wcon r , ib that ) is - is a for the ries ( to give too smidl a dency , is at omes pOllOUllced is down to s1nall of is likely for qucstiou still hold - ) makes become a characteristic , and 111 the ctld the wholly fails except for an fiuite v of.r .
Mr. O. Heaviside .
[ June of Methol of Representation by 41 .
We may then adopt another method .
Thus , ( 44 ) and ( 45 ) ise from , ( 70 ) ( 71 ) With unit operand , the series is immediately integrable without any obscurity , .
The series leads to an unintelligible result .
But let the unit operand be replaced by its simple hnrmonic equivalent .
hen when is positive , which is the required resul .
We are only concerned with positive , but it is worlh noting that when is negative , method makes zero .
This is also in accordance with the analytical method , or ( 70 ) directly integrated , for we suppose the operand to start when , and to be zero for negative , which also ihen .
42 .
As regards the derived to ( 42 ) , although I ave not examined tbem thoroughly to ascertain limits within which the stlspected numerical equivalence may obtiain , I find there is a agreement ween ( 41 ) and ) when and , even with , and the convergency confined to the first of the results being ; ( 74 ) which , when .
Again , with the much larger value , we have which is a very close agreement .
is promising as regards further numerical greement when 1893 .
] On Operators in Physical ffathematics .
is made larger , but the promise not fultilled when is as big as 10 .
Take and in the series ( 41 ) , ( 42 ) , so that ( 75 ) .
( 76 ) Hore makcs a little Jess than hile he is 2965 , which is little changed by the next .
But a peculiarity in the scries .
The .
from ency to at the term is so mmensely rapid this fact alone might render series quite unsuitable for pproximate numet.ical calculalion .
A portion of the .
the term might be ( in the last examplc ) , but when this term is a large multiple of the least , no definite ation is obtainable .
What is the of alence 7 kefch of .
zent of alence and ies ( ?
to S 49 ) .
43 .
In the preceding , I have posely avoided definition of " " equi } nce ] .
in exnmple raf her ecept , I have to ] he , and the method of them , speak for .
Besides } , I could not a satisctory definition hich I could fecl sure would not quire subsequent revision .
lIathcmatics is an experimental , and detinili do not come first , but later on .
make when the uatnre of subject has developed itsel It old bo absurd to lay the eforel ) .
Perhnps , therefure , the best thing I can do is describe successive es of ating to alent : approximately representative of personal experience . .
Complete noran c ( b ) .
A convergent has limit , and therefore a definite value .
A divergent scrics , on the contrary , is of hnite value , of course .
So all solutions of physical musb be linite tcrm or in convcrgent series .
vise nonsense is made .
The Use of ( Scries .
) of Continuous Sc es .
44 . .
-opening .
But in physical problems ser es are actually used for calculation .
A otable e is Stokes 's divergent formula for the oscillating function .
lfo showed that the error was less than the Iasb iucluded .
Nfr .
O. Heaviside .
[ June 15 , eries of this kind have the terms alternately positive and negative .
This seems to give a clue to the numerical meaning .
The terms get and bigger , but the alternation of sign prevents the assumption of an infiuite value , either positive or negative .
It is possible to imagine a quantity split up into parts alternately positive and ative , and of ively i magnitude ( after a certain point , for example ) .
It is a bad ement of parts , certainly , but understandable roughly by the initial ence .
So the use of alternating divergent series may be jusvified by numerical convenience in approximate calculation of the the But , by the reasoning , a direct divergent series , with all terms of one sign , is of infinite value , and therefors out of court .
It cannot have a finite value , and cannot be the solution of a physical problem involving iinite values .
This seems to be what Boole mean in his remalk on p. 475 of his ' Differential Equations ' ( 3rd edition):\mdash ; " " It is known that in the employment of divergent series important distinction exists between the cases in which the terms of the series are ultimately all positive , and alternately positive and ative .
In the latter case we are , accol.ding to a law , permitted to employ that portion of the series which is eItt for the calculation of the entire value He proceeded to exemplify this by Petzval 's integrals .
The ument is equivalent to this .
Change the sign of in the Series , equation ( 3 ) above .
the result be .
Then we must use the Series when is positive , and when is negative .
This amounts to excluding the direct diverg.ent series altogether , using only the .
That is , we have one soIution , not two .
Professor Boole did not ) what the " " known law\ldquo ; was .
His above authoritative rejection of dircct divergent led me away from the for many .
The plausibility of the argoment is evident , as evident as that the value of a direct divergeut serics is infinity .
Dit'ergent Series as Differentiating Operafors .
45 . .
Later on , diverg.ent se1ies presented themselves in an entirely diifel.enb mannel .
IIL the solution of oblems b means of or anal t , the opel a themselves may be either , or alt divergen } , is , they are so when alCR , entiator regarded as a quaIltity .
the dicated by operator are carried out upon a ction of tlIe of the arises , and in a form .
we of the at last .
It Ilulle a its involved .
But analytically considered , of divergency does not arise .
he } ) opcr use of ent 1893 .
] On Operafors in Ph ?
siced llfathematics .
anal.ytical ators to obtain ] gebraical solutions .
The series and above refer ] to then truly the two independent solutions of a differential equation , and neither should be , for they are uatural companions .
Disappearance of th Direct DiverfSeries .
46 . .
But , still .
the subject along the same lines , is view is sooll found to be perfect .
a to a collvergent solutio1 ] may lead to a solution by another .
Or it may lead to the same algebrnical by se .
These and other considerations show that dive.gent series , even ss hen colltinuously , must be considered numerically as ebraically and ytically .
But in the analytical use of a direct or continuously di ent scries every term must be used , if the resuIt is a series .
Yet it plain that callnot the whole en series , because it has ) limit .
And on examination we that the initial convergcnb patt of the continuously series gives the of the iunction in the same as an divergent .
In the lat. case we come nearest to the value by stopping at smallest term , the oscillation is least .
If we now makc all terms positive , so that the series is contintlolls[y di , and treat it in way , and stop when addition made by a is the smallest , we come near the We now seem to have something lik-e a disfillct theory of divergent selies .
The supposed disbinction between the and the colltinuous ( series has ) ticnl equivalence , one gent , , may all in dive one to be nted .
alence exists also , is erned by the initial eucy .
and Deeper the Calculus .
xlytical , , and lences .
47 .
( f ) .
The last view is a distinct advance , and it is ] in the case of many equivalences , lclnding some are of in mathematical .
But , exalll tion stlows t the last word has not been said .
on to explain the meaning and origin of equivalent serics , led to of alized dvolving the vorse f as a completely continuons both to methods of equivaleut forms to aIly e , and in a generalized mannei , previous mples being meroly special extreme of Mr. .
O. [ June 15 , the general results .
We also come to confirm the idea we have recognized that alence may be understood in three distinct senses .
, aloebraicalb .
, and numerical .
The first use made by me of equivalent series , one of which is continuously divergent , ; analytical only .
The second use was numerical .
The third is algebraical , through the generalized algebraical theorems .
We also see that equivalence does not necessarily or usually mean Thus the eries A are analytically , algebraically , and numerically equivalent with positive .
Bot they are ) aIgebraicalIy tical .
The identity is given by .
This point is rather important in some transformatio1ls , and explains some previously ine.xplicable peculiarities .
Thus , the series A is real whether be real or a pure imaginary .
In the latter case , we get the oscillating function , the Fourier cylinder function .
But the equivalent series becomes complex by the same transformation .
The above-mentioned identity explains it .
The second solution of the oscillating kind is brought in , as will appear a little later ( S 70 ) .
Partial Far.lure of Interpretation of Numerical of Divergent Series by lvergence .
Furthcr Explanation yet required .
48 .
( g ) .
But whilst we thus greatly extend our views concerning serics , the question of numerical equivalence , which just now in ( f ) seemed to be about settled , }ecomes again obscured .
The property the value of a divergent series , including the continuously may be estimated by the itially convergent part , a very valuable .
But the property is not generally true , and , in fact , sometimes in a vel'ymarked manner .
We must , therefore , reserve for the present the question of numerical equivalence in general , and let the explanation olve itself in course time .
If definitely andable numerical equivalence of series were imperative under all circumslances , then I am afi.aid that the study of the subject would be of doubtful value .
But the matter has this limited range , a very important ) ] icatJon of divergent series boing their use , which is from the numelical difficulty .
example , the extreme forms of the binomial theolem may , when considered numeric , ally equivalent , be utterly useless .
Yet they may be employed to lead to other , either convergent , , or it may be divergent , but with a satisfactory initial gence colltrasting with the original .
Note that the series may sometimes ts ke the form of definite of infinite or of indefinite value .
In any case we should not be misled by apparent unint , eJligibiliy to ignore ) subject .
That is not the way to get on .
We have seen the error fallen into by Boole and others on the subject of divergent series .
It is not so long ago , , since 1893 .
] erators in Iathematics .
of the hesb 1epute could not see the validity of estigabions based upon the use of the algebraic imaginary .
The resuIts reached were , according to them , to } regarded as estive merely , and equired proof by methods not involving the .
But familiarity has bred contempt , and at the present day the ginary is a generally used erful e , which I should think mosb ) naticians consider can be trusted if well to give valid proofs , it does need cautions treatment sometimes , and perhaps auxiliary aid .
* Applica ) to a for 49 .
Let us now pass on to view the ] in its generalized nspect .
One way of genelalizing is to regard it as the limit of .
Now , using the ized binomial theorem ' here any value and the step is unity , we obtain by this process ( 78 ) here the acccnt means differentiation to ? ?
, after which the special values are given to the lment .
Or , since , and , ( 79 ) if is the inverse factoriaJ , therefore .
( 80 ) But also by S 17 and equation ( 94 ) , Part I. So we reduce to .
( 81 ) * Perhaps we fairly regard the theory of lized { lysis as being now the sanlP stage of development as the theory of the was opmcnf of the modern theory of functions .
Not I know bout t tter ; the big book lately turned out by Forsyth to me quite unexpecbed elopments .
126 Mr. O. Heaviside .
[ June 15 , 50 .
To obtain the common formula for the logarithm , take Then , since , \amp ; c. , we reduce ( 81 ) to . .
( 82 ) When in ( 81 ) , we have .
( 83 ) Now here all the differential coefficients of the inverse factorials may be put in terms of by means of the formula , ( 84 ) which follows from ; but since the resulbing formula does not ssem to be usefnl , is complicated , it need not be given here .
Deduction of ormula for 51 .
If we diffelentiate ( 81 ) with respect to , we obtain , where the second form of serics is got ) increasing by unity in the first .
Here note that we have a definite expansion , wher !
in S 32 we found the binomial ) ansion to be indeternlioate .
lVhen in ( 86 ) we have , of cotlrse , the special . . .
It is also right when Deduction of for 52 .
Now regard ( 86 ) as truc ally , and we can a formula for , first put for ing 1893 .
] On erators in Iathematic ing , we obtain .
This is quite correct when , when we obtain the ordinary fornlula . .
Anot , form of ( 88 ) is .
( 89 ) Now when , the square of the sine equals uuiby throughout , ivino . . .
( 90 ) Since we also have , ( 91 ) the product of ( 90 ) and ( 91 ) should be .
That is , ( 92 ) Going by principles of he of series , we sllou Id concl ( lde that the coefficient of \ldquo ; was , and he coeflicients of the powers of were .
But is genelally in scl.ies of he prcsent , as we have eady xcrnI)fied .
Thereforc , to sce it goes in the have calculated the value of the coefficicut of .
By ( 84 ) l1avs Mr. O. Heaviside .
this we may derive , when is a positive integer , ' ( 94 ) and also .
( 95 ) Therefore , when is positively integral , we have which makes the coefficienb of in ( 92 ) become here , for brevity , stands for .
It is readily seen that the complete coeffioient of vauishes , and the remainder reduces to Therefore the coefficient of in ( 92 ) contributes one-half of the total , the other half be given by ( or rather , be equivalent to ) ) sum of the terms involving .
Although I have not thoroughly investigated this , there did not appear to be any consistency .
on Equivalences in Factorial ) .
Verificutions .
53 .
If it is given that , ( 98 ) it does not , as already remarkel in effect , follow that is a definitely unique nction of .
But it is somebimes true , and then the equation ( 99 ) may require the vanishing of every coefficient .
For example , using ( 88 ) above , if we differentiate it to we obtain ( 100 ) 1893 .
] On erators in ysical Jnatics .
elore , by adding this equation to ( 88 ) , obtain .
( 101 ) Now it is a facb that bhis is true , by term , , 1 , 2 , 3 , \amp ; c. Bnt ( 101 ) is not in the same manner generally .
Only when , that is , when is a negative , do } have , ( 102 ) by ( 93 ) , which is general .
Put to suit ( 101 ) .
, ( 103 ) by .
Therefore ( 101 ) is the same as , ( 104 which does not vanish term by term , except for the special values of indicated .
rating ( , we obtain constant .
( 105 ) 54 .
case we have already had , when the constant is 1 , so it should bc 1 gellelally .
The case !
is rpesented by , ( 106 ) and the following is a verification : is .
( 107 ) validity of this process of may be doubted , there no iuconsistency exhibited .
55 .
The other formula of a similar kind , , ( 108 ) VOL. Mr. O. Heaviside .
[ June 15 , similarly treated , gives ( 109 ) That is , , which is a case of Taylor 's theorem , if we do not go too cIose to the boundary where the operand begins .
That is , regarding the operand as , it is turned to .
Application of Exponential to other Generalized Formuloe intolving the Logarithm .
56 .
Now return to the fundamental exponential formula , ( 110 ) and derive from it some other logarithmic formulae .
Differentiate to , then .
( 111 ) A second tion to gives .
( 112 ) differentiation gives , ( 113 ) and so on .
Or , all together , -r . .
( 114 ) Now combine them to see if they fit .
, we have the eIementary mula ( 115 ) and , by the use of ( 114 ) , becomes , ( 116 ) which , by Taylor 's theorem , is the same as fff , ( 117 ) required .
57 .
Again , differentiate ( 111 ) to .
We obtain , ( 118 ) by using ( 111 ) , and So 1893 .
] On erators in matics . .
( 119 ) Here the ctor of is identical with ) , by ( 84 ) , which corroborates .
58 .
Returning to ( 111 ) , if we try to make a series for in powers of we obtain Tbis is done by making be the represenliative power hout , by reducing the value of by unity in the second term in the first series , by two in the third term , and so on .
Or ( 122 ) This is striking , bnt not usable .
Also , if we try to get a series for we fail .
The property ( 84 ) comes in , and brings us to in the end .
This failure is not obvious a priori in factorial mathematics .
Deduction of a Logarithmic 59 .
Now let the formula ( 111 ) be specialized by taking .
We then have .
( 123 ) Here , for the negative values of we have ( 124 ) and so on , whilst for the positive we have , O. Heaviside .
[ June 15 , by ( 84 ) .
Using these in we obtain ( 12b ) The first series is the ordinary expression for with the terms inverted , whilst the latter coMains a reminiscence of the companion to the Fourier cylinder function . .
To see whether there is a notable convergency for calcnlation , take Then This is evidently by the look of it , especially when represented .
Also So ( 126 ) gives By comnlon logarithmic tables we find .
The difference is .
Doing it another way , we may prove that , ( 127 ) is an infcresting mation .
This , with , gives 13203 , and produces a much closer reement .
It is probably fortuitous .
of the Last .
61 .
We can establish ( 126 ) iudependently thus:\mdash ; We have ( 129 ) 1893 .
] On fors in fics .
rate to .
Tbeu .
, ( 130 ) when is some constaut introduced by the ation .
To find it , note that the series with the exponenbial factor vanishes when is infiuite ; so ( 130 ) gives , with .
( 131 ) It is not immediately obvions that the function preceding th ) arithm in ( 181 ) increases inlinitely with .
Bnt by ( 127 ) we may regard it as the ratio \mdash ; and we see that the terms in the numerator become infi than those to correspond in the denominator .
A Formula for Euler 's Constant .
62 .
Next examine whether ( 131 ) give a rapid approximation to of C. When we get When we get Vhen x we get So with the error is about 5 only .
The usual mula , with is very slow .
Ten terms make .
Twenty about hich is still far wrong .
We seo that ( 131 ) will with a moderate ] of 63 .
In passing , we may nobe that the functiou is reprcsented by , ( 1.34 ) .
O. Heaviside .
[ June 15 , and also by , ( 136 ) which may be ussful later .
Deduction of Second Kind of Bessel Funcbion , , from the Formula of the First Kind .
64 .
A similar treatment of the generalized formuIa for the FourierBess function leads to the companion function .
Thus , take as in ( 76 ) Part I , the value of being .
Differentiate to Then .
( 138 ) Here ake the special case .
Then we have The third line is apparently zero .
But it must , as we shall see , be retained , though in a changed form .
Or Iog .
( 140 ) Another way .
Automoti 65 .
Now the right member is certainly not zero , for it the companion of , as may be proved in various ways , classic ) nd unclassical .
One way is from the formula for thus , .
( 141 ) On Operators in Fhysical Jfathematics .
When is not an integer , and are different , and represent two indepelldent soIutions of the characteristic differential equation .
But when is any integer , positive or negative , they become identical , so only one solution is got .
another is ( when ) represented by the rate of iation of with when hus , ' ( 14 ) which , when , is by inspecbion the function on the right side of ( 140 ) .
Notice that this method of obtaining the second solution , like the just preceding method , gives it immediately in the form propel .
lie standardized so as to vanish at .
The constant comes in automatically , and requires no separate evaluation .
The Operator producing .
66 .
But our immediate object of attelltion should be the fun.tion on the left side of ( 140 ) .
How it can be equivalent to the member is a nlystery .
It is cerbainly an extreme form , if correc.t .
We may write it in the form , ( 143 ) is .
Now the other function is , ( 144 ) itllout a mystery , and we sue at once that thsse forms are aua , to the latter , correspondino to ( 144 ) , being obvious , whilst the analogous to ( 14:3 ) , is all considered exsee tions ( , ( 72 ) .
The unintel l of evidence of its .
More puzzling than it have becll cleared up .
67 .
We also employ the , of had parate vions .
it and rite fur 1,36 : O. Heaviside .
June 1 , ( 147 ) where we see that -he operator ( 143 ) appears .
Integrating , we have , ( 148 ) comI ) which with ( 140 ) .
we see that ; for which a tion would be desirable .
Componion , dericed from panion Operators , expressed in Descen ) Series .
Also in Ascendnng Series .
68 .
Passing , however , at present to mors operators involving the two solntions and differont forms thereof , it will be convenient to introduce a notation and standardizstion shall exhibit the symmetry of relations most clearly .
Thus , Jet ; .
( 151 ) Here is a constant and is .
Superficially considered , ] functio1ls only differ in one being times the other .
But the common theory of the inary does not hold good here , or in opel.ators geIlerally .
In a series we have ( 152 ) as already shown .
It is twice the function , equation ( 3 ) .
, we may rate ( .
Introduce the factor , thus , ( 15.3 ) 1893 .
] On ( otors in Expand in of , and then integrate ; then ( l5 5 ) Thus the function ( qx ) only differs from in the changed , except undel the radical .
These are most primitive solutions of the characteristic equation , and are usefnl as operators relating to inward and outward going cylindrical waves , as well as for numerical poses .
The function is also expressed by ( qx ) .
( 156 ) By here and later should be understood merely the ascending . .
to the Oscillating iFunctioJlS and , both Ascending Series .
69 .
The connection between these functions and and the tory functions is very important , but was in one respect exceedingly obscure to me ulltil lately .
Thus ( 157 ) and ( 156 ) usually reckoned to be , companion solutions ( unless as regards the numerical factor ) .
But if we take in ( 157 ) , the function remains real , and becomes the oscillatory function , the linder function of Fourie .
Thus . .
( 158 ) On the other hand , the same transformation in makcs it complcx , on account of the logarithm .
Thus , using in the well-kuown mula for we convert ( 156 ) to $ Mr. O. Heaviside .
[ June where is the same as in ( 158 ) , and is its oscillatory companion given ( 161 ) What is obscure here is the getting of only one oscillating function , and of two from .
In corresponding forms of the first and second solutions we should ) both oscillating soIutions to arise in bh cases .
However this be , the transformation ( 160 ) is in agreement with the other form ( 155 ) .
For , if we make the change in it , we obtaill the same formula , provided and a given by ( sin-cos ) , ( 162 ) \mdash ; si o , .
( 163 ) where and the real functions of given by .
Now here ( 162 ) is Stokes 's for , known to be eq ni1 alent to ( 158 ) .
And ( 163 ) shows that this kind of formula for the oscillating functions allows us to obtain the second soIution from first by the of to and to .
The function ( sx ) of ( 163 ) may be shown to be equivalent to the of ( 161 ) by other means , and certainly ihcations aable , because transformations the square root of the im ginary are sometimes rnxafion fto same and .
tion as regnrds the changed form of the function of ( 152 ) , there is a real and once appal.ently insurmountable difficulty .
changed the sign of thd from in ' Pnpers ' , in order to iCivs at the origin .
1893 .
] On Operators in Physical fathematics .
We knolv that and are equivalent , both anal tically and numerically .
Why , then , does the first become complex , whilst the second remains real when we take They cannot be both in changed form .
Thus ( 152 ) becomes ( doing it in detail ) ( cossin ) ( cossin ) .
( 166 ) is , the fnnctions ( 162 ) , ( 163 ) again , we have the transformation whereas becomes .
This was formerly a perfect mystery , indicative of an imperfection in the theory of the Bessel functions .
But the reader who has gone through Part 1 and SS 27 , of Part II will havs little trouble in understanding the meaning of ( 167 ) .
The functions and , though equivalent ( with pobitive argument ) , are not aIgebraically identical .
To identity we equire to use a second equivalent form , so that , as in S 28 , In this form we may take , and still have agreement in the changed lorm .
We obtain the relation ( 167 ) , provided that As I mentioned before in S22 , this formula for may be deduced from formulae in Lord Rayleigh 's ' Sound , ' derived by a due to Lipschitz , which investigation , however , I lind it rather difficuIt to follow .
We have , therefore , three principal forms of the first soIution with real and positive , , and the intermediate fornl ( 168 ) .
We have also three forms of the oscillatory functiol .
, , ( 161 ) , ( 163 ) , and ( 169 ) .
But we have only employed two .
O. Heaviside .
[ June 15 , of , and two of , in oq and harmomziug the previous thrse forms .
It would therefore appear probable that there is an additional principal formula for , and another for , not yet investigated .
Companion Funcfions .
71 .
The conjugate property of the oscillating functions is ( ) , ( 170 ) using the pair ( 162 ) , ( 163 ) , or the pair ( 158 ) , ( 161 ) .
And , similarly , ( ) .
( 171 ) Bnt , in the transition from ( 171 ) to by the relation , it is indifferent whether we take , or else .
This conjugate property is of some importance in the treatment of cylindrical problems by the operators .
Operators with two Differentiators eading to and and their Mutual Connecfions compactly in reference to lindrical Waves .
72 .
The fundamental mutual relations of and are exhibited concisely in the following , .
operators containing two differentiators , say and and .
( 172 ) Here it should be understood that either or may be passi , when it may be regarded as a constant .
But when both are active , there are two independent operands , one for and the other for .
In a cylinder problem reIating to elastic waves , we may regard as being , where is distance from the axis , and as ( vt ) , is the time , and the speed of propagation .
We have Operators in ysical Jnatics .
[ ] . . . .
, . . . . . .
( ) , . . . .
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
, . . .
lvlJere the letters in square brackets for the rpose of concisc ence .
imilarly , we have this other set , [ Q ] . . . .
, . . .
[ A ] 1 , . .
[ B ] , . . . .
[ C ] .
( 174 ) , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . .
[ D ] , . . .
[ E ] The first set is usually , though not essentially , concel.ned with an inward-going , and the second set with an ) .
wave .
exchange of and vt and of and one set to the other , so that the proof of one set proves the other .
Tn obtaiuing ] from [ P ] we egard q as a constant , or at ally rate , as passive for the time , expand [ P ] in desccnding powel .
S of and integrate directly with the resuIt , as in S , equations ( 28 ) , ( 29 ) .
To obtain , introduce the factor to [ P ] , and expand the transformed opel.ator in descending powe1s of , as in S 14 , equatious , ( 31 ) .
To obtain , we make passive , and introduce the ctor c Then expand the transformed operator powers of and integrate as in S68 , eqautions ( 153 ) , ( 155 ) ( only there the operator is , making the case [ C ] ) .
On erators in ] [ June 15 , Defails concerning the 73 .
As regards , it may be obtained from , or These have not yet been done , so a little detail is now given .
Thus , from to In the first line we expand the function ; to get bhe seoond line we integrate with unit operand ; and , finally , let operate to get ( 175 ) .
74 .
Next , from to , ( 176 ) which needs no ] aIJation , as the course is similar to the previous leading to ( 175 ) .
75 .
As regards deriving from , this may be done by Larmonic decomposition , thus , , ( 177 ) the value of which is known to be ( 175 ) .
Conversely , we may evaluate the definite integral by turning it to the analytical form .
which may be done by , and then egrating hrough the equivalcnb operator .
But this definite integral is only one of sevelaJ that may be immediately derived from the operators in ( 17.3 ) , ( 174 ) by decomposition , and it will be more convenient to consider them separately in later sections along with applications and extensions of the preceding .
On the Failure of a in Ph ) 14.3 Etastic .
Wave with corresponding Dijfusive Ware throuthe Operators .
76 .
The formulae and [ E ] are of a somewhat different kind , since the operand is the reciprocal of the independent variable .
They are proved at once by carrying out the differentiations .
Thus , for [ E ] , .
( 178 ) So , by [ C ] and [ E ] we have ( ) .
( 179 ) There is an interesting analogue to this transformation from to occurring in the theory of pure sion .
Change the meaning of from to 1 , that is , to its square root .
Then we shall have .
( 180 ) The quantity is no a velocity , bowPver .
In the theory of heat diffusion it is the ratio of the conductivity to the capacity .
This example belongs to cylindrical diffusion , and is only put here to compare with the preceding example , which belongs to the corresponding problem with elastic waves without local dissipation .
IX .
" " On a of the Law in Photography that when the Products of the Intensity of the Light acting and of the Time of Exposure are Equal , Equal Amollllts of Chemical Action will be produced By Captain DE )A C.B. , F.B.S. Beceived Jume , 1893 .
as been generally assumed tb at when the products of intensity of ight acting on a sensitive surface and the time of expobure aro similar amounts of chemical actiou are produced , with the rdinary exposures and intensities of light employed sucl ] , no doubt practically the case , and any methods of cment hitherto practicable have been insufficiently delicate to discover an departure from this law , if such departure existed .
In some recent experiments

