]\gt ; Mr. W. Steadman Aldis .
On the Numer.icat dffierence in stature was found to be inches , and the standard deviation of its distribution inches .
The correlation between difference in stature and size of family was , or greater fertility appears associated with small differences .
The observations , however , are so few ( 205 ) that the probabls error of the correlation is .
, and thus no stress can be laid on this result .
If the reader asks why is not the result in S7 conclusive , the answer must be , it old be conclusive , if the means of the husbands and wives weighted with their fertility were the same as when they were unweighted ; increased correlation would then necessarily connote that fertility was associated with homogamy .
Actually the fact that absolutely taller mothers are the more fertile alters the centre of the correlation table , and somewhat obscures the issue as to whether the whole increase of correlation is really due to homogamy being correlated with fertility .
That in man , whether from conscious or unconscious sexual selection , there is far more homogamy than has hitherto been supposed , my family data cards amply demonstrate .
If in man , then with great probability we can consider it to exist in other forms of life .
But the existence of such homogamy is of immense importance for the problem of differentiation .
The present statistics do not enable us to say whether homogamy in man is definitely correlated with fertility ; they do show that fertility is not a random character , but depends upon the relative size of husband and wife , and thus bring evidence in favour of.genetic selection .
I can conceive no more valuable investigation than a series of experiments or measurements directed to ascertaining whether homogamy is or is not correlated with fertility , but such investigation , bearing in mind Darwin 's conclusions , should carefully distinguish between exogamous and endogamous homogamy .
" " On the Numerical Computation of the Functions , and \ldquo ; By W. STEADMAN ALDIS , M.A. Communicated by Professor J. J. THOMSON , Read June 15 , 1899 .
1 .
The complete solution of the equation may be written , where ; tation of the , .
33 and , if ?
be positive integer or zero , , where --1 denoting , with the special case , and being 2 .
When real quantity , the function tscs from zero lnity , when to an infinitely large quantity , passes from zero to infinity , while decreases thy from infinity to under ihe same circumstances .
The values of the functions have been tftI ) nlated by the present writcr , and published in the ' Procecdings , ' for of .
at intervals of from to .
The elements used in the calculntion of the earlier half of these results are available for computiug the values of and in some cases when is a complex quantity .
If be a pure imaginary being a scalar , it easily seen that , where is the ordinary Bessel 's function of the first kind order .
If also denote Neumann 's function of the nth order , illd CT a function defined by the relation E. .
it can be shown without much diffioulty that . .
( 6 ) .
.3 .
The nnmerical tion of the functions to depend on that of ) and K for lues o which the convergent series ( 1 ) and ( 3 ) applicable .
In ] this it is necessary to calculate the of , tIld incidentally to compute these functions .
With the notation used in the writer 's paper on the computation of , it is easily seen that V0L .
Thns the elements , used in the computation of and , for any value of can be easily used to derive the values of and 4 .
We have further whence by addition ( 9 ) , nsing the notatio ] ] of the former paper .
Agair ] , bm ; whence , adding .
But of tha ctions G and .
Hence The portion of this ' hence 5 .
The quantities and , and the multiples of the different of have have been computed for the values of , , in the process of calculation of and , given in the writer 's former paper .
It has been , therefore , an easy ttter to find by ( 7 ) , ( 8 ) , ( 9 ) , and ( 10 ) , the , , , corresponding to the same of The former of course well known , bnt the recalculation affords verification of the correctness of the quantities .
The results given in , appended to this paper , negative lues being indicated by the use of old numeral type .
The formula used for verifying the of I and was . .
Replacing , by means of ( 4 ) , this gives . .
This fornlul has been applied throughout Table I to set of foll three places beyond thosc given .
Whcrc the last ) incrcased by unity , consequence of the first omitted ) cing c to 0 five , the is indicated a dot fter the lnst figure .
The column becll tested with tisfnctory resuIts differencing .
6 .
The value of bc readily cxpressed iu terms of tho tities , when is either zero or mity , in one or two othol beside those of lig ) or wholly .
The values of ) nlated in the Report of the British tion for 1893 , to of decimals for intervals of of unit .
Table II } the end of this gives them for the same nnmber of for the been used in the calculation of the ] ctions .
and are denoted in the ) II ) iu accordance with the notation adopted ) the Committee of the British Association , negative values being de1loted ) the use of old numeral type .
7 .
Assuming the accuracy of the yalues used for the quantities accuracy ranteed by the tests to which the Tables for I and in the former paper ) subjected , the tion between I and very check for cting i itny mistakes in ddition or copying figures in ding the of J. Thus In finding , are separately computed by ddition of late terms from and the smaller sum written down below the larger .
In cases Table I the sum of these has first been taken , and the agrcement or ( lisjgreement of this stlm the known correct value of shown either that there was no mistake , or has revealed where stlch mistake was committed , and secured its correction .
A similar test of accuracy in finding is derived from the known values of .
In like manner , since X Computation of the Functions , and .
37 the known value of , obtained in finding , gives a check on mistakes in calculating X. The known value does the same service in regard to the computation of or 8 .
By formula 8 ) Hence sin .
say ( 13 ) , hcre , the summation } ) in ffi cases from to the largest value of which gives sensible values for The values of , computed by.these formulae from the known values of , are given in Table II .
The computations evidently admit of chcck to accuracy of the ture as those given in the rticle .
Another form of the values of and is given by , reduces the computation to that of the two quantities and , so that if these be denoted by and This form admits of somewhat diflerent checks to mistakes .
The values in Table II been computed independent ] in the two so that the writer has every confidence that they be relied ou as .
The cohuun nlso ) differenced with ctory results .
: ; 8 Mr. W. Steadman Aldis .
On thNumcrical 9 .
The well-known sequence laws ( 14 ) , ( 15 ) can be utilised , the former to obtain the values of , and the latter to give a verification to some extent of the values of , by means of the formulae given in the writer 's paper on I and , which express in terms of a series of equidistant values of Thus , since replacing by , and using the values already assumed for and , it follows that Whence ' ( 16 ) .
By means of the formulae ( 18 ) , ( 19 ) , and ( 21 ) of Articles in the paper above referred to , this formula gives a check to the series of values in Table II to a considerable number of decimal places , to thirteen places with the last approximation .
10 .
For determining the values of , , the sequence law , it is convenient to denote these quantities by the symbol when is even , and by wheIl is odd .
This will be foumd to avoid irrational multipliers in the successive derivatio1ls .
Equation ( 14 ) , putting for gives The of odd and eve1l must be separately considel.ed .
First let be odd .
The eqnation gives , emembering t ' .
lVhence , if ) odd , ( 17 ) .
ntation of tl , Functions , , and .
If , secondly , be even , the equation gives , hence .
( 18 ) .
The most importftnt special cases are when .
In the first , l.ememl ) ering t ( Article 6 ) , squations ( 17 ) ( 19 ) .
In the second case ( 18 ) gives ( 20 ) .
In these derivations no labour is involved , except of addition of known quantities , and division by Mr. W. Steadman Aldis .
On the Numerical Table I. N.B.\mdash ; Negative quantities are Computation of the Functions , , and .
41 denoted by old numeral type .
VOL. LXVI .
Table II .
quantities Table II .
561790777 581466872838 OSO 738093475 393276 by old numeral type .

