]\gt ; Inbreeding in a Stable Simple with Special Reference to Cousin Marriage .
By S. M. , TACOB , I.C.S. , Biometric Laboratory , University College , London .
Communicated by Prof. Karl Pearson , F.R.S. Received Iarch \mdash ; Read , 1911 .
CONTENTS .
PAGR 1 .
Introductory 23 2 .
Brother-Sister blIarriages 24 3 .
First Cousin Marriages 26 4 .
General Sibling Families 29 6 .
Cousin riages 31 6 .
Second Cousin Marriages in Detail 7 .
Numerical Proportions of Constituents First riages 34 8 .
Percentage of Allogenic Offspring which arise from First Cousin iages 9 .
Relative Rate of Production of Allogenic Constituents by First Cousins and Non-related 38 10 .
Conclusions and Limitations 40 1 .
The Mendelian theory of the ation of unit characters , though itiis far , as yet , from being completely demonstrated , offers a simple explanation of some striking features of inheritance .
In particular , Mr. E. C. Snow has recently shown*that the gametic correlations for collaterals deducible from the Mendelian hypothesis are in close reement with the actually observed somatic correlations for man and certain other animals ; or , in other words , that a Mendelian theory of ation without dominance gives values for collateral resemblance not greatly differing from those found from observation It seems , therefore , possible that the same theory will throw some light on the problem of inbreeding , or , at any rate , will indicate to what points , on which precise data are at present , statistical enquiry should bs directed .
Without these data ) Mendelian theory cannot be corroborated or negatived by the methods of the present paper .
So far as they go , however , the statistics at present obtainable with regard to consanguinity in the parentage of albinos and deaf mutes are in approximate agreement with the calculated results , although the accuracy of the figures is too uncertain for the application of anything more than a rough criterion .
Pearson hreviously shownRoy S that this result is true for ancestral relationships , ' Roy .
Soc. Proc , vol. 81 , 1909 .
Mr. S. M. Jacob .
Inbreeding in [ Mar. 18 , The paper also gives a very simple operational notation by which the sibling families in any generation can be at once written down .
From this result the offspring of the sibling group in any generation , when these mate endogamously , can be immediately calculated .
As Mendelism , whether it be finally accepted or rejected , is likely to be a very much debated subject for some time to come , this condensed notation may serve a useful purpose in minimising braical labour .
Miss Elderton and Prof. Karl Pearson have taken a special case of Mendelian in which one of the common grandparents of the inbreeding first cousins is a pure recessive , although the occurrence of this pure recessive constituent is assumed to be excessively rare in the population at large .
If the recessive constituent is a harmful one this makes first cousin marriages appear in a very unfavourable .
As a matter of fact it will be show1l that where the recessive element is of very infrequent occurrence inbreeding greatly increases the relative chances of its appearance in the offspring of consanguineous matings as compared to its frequency in the offspring of unrelated pairs ; but at the same time the absolute frequency becomes smaller and smaller in both cases with increasing rarity of the pure recessive in the general population .
In the present paper the of the sibling group always supposed to be a representative sample of the general stable population .
The plan adopted will be , hrst of all , to obtain the offspring of brothersister and first marriages in the direct way , and then to use the operational notation to obtain the of intermarrying cousins .
Some numerical results have been calculated , and a comparison made with the available statistics .
The notation used will be the ordinary , though verbally somewhat inconvenient one in which ( AA ) , , , represent the protogenic .
hybrid , and allogenic constituents formed by a allelomorphic pair .
Where only a pair of allelomorphs is dealt with ] , the only possible stable form of Mendelian community , in there is no preferential , selective death-rate , or differential fertility , is , where the cosfficients give the frequency of the elements which they precede .
2 .
In order to obtain the ffsl ) ring of -sister we must each of the families shown in Table I on p. 41 of Mr. Snow 's paper , and , assuming that these first consist of males and females in equal 'Eugenics Laboratory ' IV , " " On the Measure of the Resemblance of First Cousins Dulau and Co. , 1907 .
1911 .
] Stable Simple Mendelian numbers , mate them in pairs , and multiply the offspring by the frequency of the particular type of family .
The total number of children in each first family is taken as 16 where does not vary from family to family .
In practice , of course , as eyery family does not consist of a fixed number of children with the brothers and sisters equally numerous , this assumption would appear to be rather a large one .
It will be necessary to consider later how far the of , and its equal partition into male and female , are hypotheses likely to give reasonable approximations .
Certainly the complexity of the analysis would be very greatly increased if possible size of family , with every possible distribution of the sexes , were considered , and probably no very great modification of the result is to be ated .
Let the offspring of each brother-sister lnarriage be , where is also constant , but may differ from 4 In the first family there are brothers and sisters .
There will ) of brother and sister , and the progeny will be 8 .
In the second type of family there are brothers ( AA ) , and brothers ( Aoe ) , and a like distribution of sisters .
The will be In the third there are brothers , and the is In the fourth type there are ( AA ) the progeny is 8 In the fifth type there are brothers , and the is In the sixth type there are ( aa ) brothers , and the progeny is .
The frequency of these types is Iultiplying the .
by frequencips and , we find the total offsprmi , of all types of brother-sister cres is the factor ( AA ) ( aa ) ( aa ) .
Mr. S. M. Jacob .
Inbreedvng in a [ Mar. 18 , Dividing by 4 ( the factor 8 is thus omitted ) , and collecting like terms , the result is Thus the progeny of brother-sister marriages is 8 multiplied by This last expression gives the relative frequency of the enic , hybrid , and allogenic constituents in the offspring .
The percentage of pure recessives in the offspring of brother-sister marriages is ) , as against a of in the offspring of unrelated pairs .
The relative frequency is the ratio of to unity .
This increases indefiniteJy with the ratio to , that is with the greater rarity of the allogenic element in the general non-inbreeding population .
This result will be seen to hold for the offspring of inbreeding couples , however distantly related .
For this reason the discussion of this point will be postponed for fuller consideration when cousin are dealt with .
The absolute frequency of the allogenic constituent in the offspring of brother-sister marriages , which approaches zero when is very large .
In the case given by Miss Elderton and Prof. Pearson this ratio is 1/ 4 .
For this value of the ratio , approximately , and the proportion of the allogenic constituent in the non-inbreeding population becomes Thus , for an evil which is recessive and obeys the Mendelian formula , its appearance in 25 per cent. of the offspring of brother-sister marriages must be compared with per cent. of patent evil in the offspring of non- inbreeding couples .
3 .
We will now turn to first cousin marriages .
Mr. Snow 's Table can be shortly written out by collecting like ternoe iri each sibling family , and the same process as that used for obtaining the offspring of brother-sister marriages can be readily applied .
The fertility of the first sibling generation when mated to similar first sibling families is taken to be 4 .
It is not necessary to give here the reduced table .
Each constituent in Mr. Snow 's Table denotes that there are families , each with members , and these members are brothers and sisters to members of the same family , and first cousins to members of other families in the group .
Thus in all there are in each group 64 individuals who are either brothers and sisters or first cousins .
1911 .
] Stable Simple ]Julation .
As before we assume an equi-partition of the sexes , that is there are to be 32 males and 32 females in each group .
Not only are males and females to be equally numerous in the group as a whole , but also in each family is the number of protogenic males to be equal to the number of females , and so on for each type of constitution .
If we number the groups in .
Snow 's table from 1 to 36 , from left to taking the rows in turn , we have different types of second sibling family .
Then , mating male and female , and putting the fertility of each mating as , the offspring of inbreeding second siblings is ( AA ) [ the expression for the coefficient of ( AA ) with and interchanged ] .
When the terms collected , this reduces to ( ) .
( 2 ) This last expression , multiplied by 16 ives the distribution children of all second siblings .
Now we have mated together at random all the second in each group of families , so that both first cousin marriages and brother-sister marriages have been included .
To obtain the offspring of first cousin marriages only we must subtract from the above result the offspring of brother-sister in the second generation .
These can be found from the expression for the children of brother-sister in the first generation by chan into into multiplying by 16 , and dividing by 16 The first three changes are clear .
The final division by 16 is necessary , because , without it , we should obtain the of second siblings , on the supposition that all these siblings contracted brother-sister marriages ; whereas , if , as has been done above , the second siblings mate at random each in their own group of families , there will be 16 cousin marriages to one brother-sister marriage , there being 16 families in each group .
Hence the children of brother-sister in the second generation are given by the expression ( 1 ) multiplied by .
Subtracting this from the expression for the offspring of all second siblings , we get , omitting the factor 2 mt ( aa ) ( ) .
( 3 ) Mr. S. M. Jacob .
Inbreeding in a [ Mar. 18 , The proportion of the allogenic constituent in the progeny of first cousin arriages is ( 4 ) excess of the recessive constituent over the amount produced by the non-inbreeding population is .
( 5 ) The ratio of the two rates of production of allogenic element by first cousin and non-consanguineous respectively , is .
( 6 ) The question now arises as to what value should be given to .
Theoretically , even in man , the possible gametes of both male and female are very large in nulnber , and the resulting zygotes would also be very numerous .
It seems not unreasonable then to take the actual population , in which each family is limited by a variety of causes , of which no account can possibly be taken , as a random sample of the population which would arise , if every possible were formed and developed .
In this case we must put practically infinite .
In doing this we do not make each lual family infinite , but we assume that existing families are a random sample of what would occur , if every possible gamete of both parents survived to form a zygote .
This , of course , would not be legitimate if applied to the individual family , but it seems reasonable as means of predicting the constitution of the whole population .
To assume that each family has a variable number\mdash ; as a rule too few to give the possible Mendelian variations\mdash ; would lead to an immense aJgebraical extension of the work , and it is hard to see that it could lead to results differing from the supposition that the actual population is a random sample of the theoretically possible population , as an indication of the variation produced in the results by limiting fertility .
I have put 16 equal also to the average size of the family met with in practice .
The chief objection to this is that it obviously ives fractional frequencies , within the individual family , to some of the Mendelian possibilities .
In the present case , to make infinite means that the expression ( 2 ) alone gives the of first cousin marriages , whereas it actually includes 1911 .
] Stable the offspring of the brother-sister marriages in the second generation.* The inclusion of these marriages has the effect of increasing the proportion of and allogenic element at the expense of the hybrid constituent .
Seeing , then , that the proportion of allogenic element in the offspring of first cousins increases with the fertility of the first generation , where the fertility does not vary from family to family , it would appear not unlikely that where the fertility is made variable from mating to , the proportion of allogenic elements will not oreater than the value determined by putting the fertility constant and equal to the maximum value recorded in the observations .
Further , it would appear that the true value must be greater than that given by the least recorded fertility .
lt is true that the gametic correlations of Mendelian collaterals , by putting infinite , agree closely with statistical results for somatic characters , become too small when we put 16 , say .
We have , in fact , the values , for the fraternal , avuncular , and first cousiu ametic correlations respectively .
However , until Mendelism is more firmly established , this cx-post-facto justification , of some , cannot be arded as decisive .
In the numerical calculations that follow , the results have been iveu for infinite and for Before considering , however , the numerical consequences of the results obtained for hrst cousin , the general expression for the arising from the intermarriage of rade of cousin will be ated .
4 .
If we take two populations of the forms ( aa ) and ( aa ) , where , and mate them at random in every possible way , the offspring of each mating , the result is easily shown to be where , after multiplication , ( AA ) is written for for Aa , and ( aa ) for .
Hence the rule for finding the offspring of any two populations is easily seen .
We simply write down for each population the number of each kind of allelomorph add .
We thus obtain two factors which , algebraically multiplied , give the required resultant offspring , provided , Aa , , are interpreted as noted above .
Thus , in the first of the two populations given above , the numbers of dominant and recessive allelomorphs are respectively and , and these , multiplied by A and respectively , give the first of the required factors .
* This of course indicates that brother-sister marriages , when all matings are random , .
would only form an indefiuitely small fraction of cousin marriages .
Mr. S. M. Jacob .
Inbreeding in a [ Mar. 18 , Disregarding for the present the absolute fertilities in each generation , we see that the families of brethren or first siblings can be written in the symbolic form where the ordinary algebraical operations are first carried out , and then interpreted in the way noted .
Similarly , it is easily verified that the second sibling groups of families can be written symbolically in the form Let us assume that this result is generally true .
We would then have the nth sibling groups given by .
( 7 ) We will prove that if this result holds for the nth siblings , it will be true for the siblings , and thus , inductively , that it be universally true .
For brevity write First of all the types of families in the generation will be obtained , and then their frequencies .
The eneral type of family in the generation is given by where and have any values from to This gives a group of type ( aa ) .
This family group written in operational form is\mdash ; where can take any value from to 2 Thus the sibling group will be formed of types of family produced by two operators of the form where has any value from to 2 Thus the types of family groups in the sibling generation are given by 1911 .
] Stable Simple Mendelian Population .
We shall now determine the frequencies .
From 7 ) we see that the frequency of the general type of family group in the nth sibling generation is given by the product where and can take any value from to Now , whenever , we get an operator of the form Hence the frequency of this type of operator is where .
Now take the identity and equate the coefficients of on both sides .
We have at once Hence the frequency of the operator of type is This is the operator which operating on itself produces the ibling families .
Hence the generation is given symbolically by .
( 8 ) Thus it is shown that if the result ( 7 ) be true for the nth sibling generation , it is true for the .
But it has been verified for the first and second sibling generation .
Hence it is universally true .
5 .
To obtain the offspring when the nth generation of siblings inbreeds , we take each of the families in ( 7 ) and mate each type with itself .
The result will be given by taking in ( 8 ) only the square terms , and omitting the cross-products , except that the frequencies are not to be squared , as we divide each group into an equal number of males and females , as has been done throughout .
We have then , absolute frequency , the offspring of sibling groups given by Consider first the coefficient of ( AA ) .
It is Mr. S. M. Jacob .
Inbreeding in a [ Mar. 18 , Now let us evaluate Thus Take next the coefficient of ( Aa ) .
It is The coefficient of is clearly found by interchanging and in the coefficient of ( AA ) , and it is Thus the distribution of the constituents in the offspring of inbreeding nth siblings is ( after dividing by 2 ) , Hence the offspring of nth cousins is ( aa ) .
This , of course , includes the offspring of brother-sister , first cousin , second cousin , up to cousin marriages .
Thus the proportion of the allogenic element determined by ( 9 ) is .
as against in the non-inbreeding population .
1911 .
] S'table Simple Mendelian The ratio of the two quantities of allogenic constituent is which agrees for with the result already obtained for brother-sister and first cousin marriages .
6 .
We will now consider in more detail the of intermarrying second cousins , introducing the necessary fertility factors and the deductions for the inclusion of brother-sister and first cousin marriages in the result already obtained .
The distribution of all the offspring of second cousins , first cousins in the second generation , and of brother-sister lnarriages in the third generation is found by putting in ( 9 ) , and we have ( 10 ) The absolute frequency of each type must now be considered .
We started with individuals who mated in every possible way with other individuals , and had to each Hence the first siblings number The ext fertility being 4 , the second siblings number The next fertility being 4 , the third siblings number 256 Let the fertility of third siblings inbreeding ith each other be .
Then the children of third siblings will number 1024 Now the number of individuals in ( 10 ) is .
Hence ( 10 ) must lee multiplied by to give the absolute frequencies .
Now in the expression ( 2 ) multiplied by 16 we have the of all first cousin and brother-sister marriages in the second eneration .
We must transfer this to obtain the offspring of first cousins in the second generation and of brother-sister marriages in the third generation , and subtract the result from the offspring of third siblings generally .
The result is obtained by changing into 4 into into and multiplying by , and finally by 16 , since the chance of a brother-sister and first cousin marriage in the random of third siblings generally is of the whole of Hence the offspring of first cousin in second eneration and of brother-sister marriaoeso in the third neration is 64 multiplied by the expression ( 2 ) .
That is , we must subtract ifrom the expression for the offspring of third eIlerally .
VOL. LXXXIy.\mdash ; B. Mr. S. M. Jacob .
Inbreeding in [ Mar. 18 , Omitting the common factor 4 tmn , the result is Thus the offspring of second cousin marriages is given by .
The proportion of allogenic element is and the ratio of this to the amount of allogenic element in the is For this becomes , and for 16 it is , the onding values for first cousin marriage being and 7 .
We shall now proceed to consider the results obtained with particular reference to the marriage of first cousins .
In the first place it will be necessary to examine how far the numerical proportions of pure dorninant , hybrid , and pure recessive differ in the offspring of first cousins from the proportions in a non-inbreeding population .
* For this purpose the ratio of to has been selected to in the llon-inbreeding population 25 , 10 , 5 , 4 , 3 , 2 , and 1 per cent. and the fractions 1/ 1,000 , 1/ 10,000 , 1/ 20,000 , and 1/ 1,000,000 of pure recessive .
Table I gives the percentages of each type of constitution in the offspring of non-consanguineous and of consanguineous first cousin marriages .
This table illustrates the absolute decrease in the percentage of pure recessive in the of consanguineous , while it is seen that the ratio of the percentages of enic constitution steadily rises as the frequency of occurrence of .
the allogenic element diminishes .
If we have an evil which is a Mendelian recessiv'e and is of common type , such as tuberculosis , then first cousin will not be much more likely to produce a defective offspring than any other kind of By a " " reeding population\ldquo ; in this paper is meant one in which inbreeding is not universal , i.e. , it corresponds to the general population resulting from random mating .
It is not intended to imply that tuberculosis has been proved to depend on a simple Mendelian recessive factor .
1911 .
] Stable Simple ]Iendeliajl Population .
Table I. It is only for a rarely occurring evil which is recessive that uinity will probably havs a marked effect .
8 .
Prof. Pearson has suggested to me that in order to compare the results obtained with statistical data , it is necessary to determine on the Mendelian hypothesis what percentage of pure recessive individuals , in a population in which there are both consanlguineos and non-consanguineous marriages , is the offspring of related parents .
For the application of this test it is necessary to know what is the frequency with which cousin marriages occur .
This , however , is Mr. S. M. Jacob .
breedmg in a [ Mar. 18 , not known with any of certainty , various determinations which have been given ranging in value from as much as 1 to 8 per cent. of the total number of marriages .
It is beyond the scope of this paper to consider the accuracy of these determinations , and they do not claim to be more than approximations .
So far as can be judged , however , a first cousin rate of 5 per cent. such as may occur perhaps among the , is not likely to be exceeded in any considerable group of the population , though it may apparently be as small as or per cent. for parts of the community of low social status .
Accordingly , in what follows the determinations have been made on the basis of rates of first cousin marriage from 1 to 5 .
For the moment , only first cousin marriages are dealt with , all othel marriages assumed to be non-consanguineous .
Let .
be the of first cousin marriages .
Let be the proportion of the offspring of first cousin marriages which has the allogenic constitution .
The value of is given by the preceding analysis .
Let and be the proportions of allogenic constituent in the offspring of non-related couples , and in the population at large .
As a rule , will be the datum available from statistics .
Then , clearly , the total proportion of allogenic element in the whole population can be expressed in two ways , which together give rise to the equation Now , if we put , we have and ' ' where Thus and .
( 12 ) This is a quadratic which determines , and thus the values of and can at once be obtained .
The percentage of the allogenic population which is the offspring of first cousin marriage is then given by .
The value of this function has been calculated for frequencies of occurrence of the aUogenic constituent at chosen intervals from 25 per cent. of the whole population to one case in a million .
The resuIts are ooiven in Table II .
Here , again , we observe that , when the recessive element is a very rare 1911 .
] Stable Simple ilIendelian Population .
one , the likelihood of its possessors being the offspring of first cousins is much greater than the frequency of first cousin marriages would lead us to expect .
The statistics available for the purpose of comparison with these results are unfortunately extremely meagre , and they cannot be as more than approximately accurate .
Miss R Mi Elderton has kindly placed at my disposal her to the percentage frequency of consanguinity among the parents of albinos .
Miss Elderton surmises that , when in her statistics it is not distinctly recorded that there was no consanguinity between the parents , it is possible that there may have been some relationship .
Hence she obtains two results : firstly , by including all available and , secondly , by excluding those families in which the presence or absence of consanguinity in the parentage is not distinctly recorded .
the first case she finds that 14 per cent. , and in the second that , , of albinos are the of first cousin and double cousin marriages .
Now , albinism is usually regarded as from a Mendelian factor , although strict proof of this is as yet that it is such a Mendelian unit character , the above results can be applied to it .
Prof. Pearson estimates that albinism occurs in about 1 individual in this frequency , we see from Table II that if the general rate of first cousin marriage is about 1 or 2 per cent. , the percentage of albinoe that arises from first cousin marriages , as found from calculation , is of about the same magnitude as that found from direct statistics .
If the , of first cousins form a greater proportion than 1 or 2 per cent. of all , then the calculated result becomes too great .
It is , however , quite impossible as yet to say definitely whether the percentage of first cousin marriages largely exceeds 2 or not .
In any case the approximate character of the other data precludes us from obtaining more thau a rough test as to whether the calculated and observed results are of more or less of the same order of magnitude .
The only other numerical test obtainable has been given by the data as to deaf-mutism .
In the American Census of 1890 it was found that there were enitally deaf persons in 10,000,000 inhabitants , or , about 1 person in .
It is probable that most of these were deaf mutes .
Miss lderton finds that 8 per cent. of deaf mutes are t , he offspring of first cousin parents .
The calculated result for 1 per cent. of cousin marriage lies between 88 and for 16 , and between and for Here again the calculated result to be somewhat too high .
Both in this case , however , and in that of albinism it may very well be that there is Mr. S. M. Jacob .
Inbreeding in [ Mar. 18 , a selective infantile death rate , unfavourable to albinos and deaf mutes , and that , consequently , the conditions of albinism and deaf-mutism are really of more frequent occurrence than would appear from a count of the population at any single instant .
If this is so , Table II shows that there will be a great reduction in the calculated number of deaf mute and albinotic persons who are the offspring of a consanguineous parentage .
Another point to be noticed is that Table II has been formed on the assumption that first cousin marriage alone has a greater rate of production of allogenic characters than the non-inbreeding population , whereas some of the excess rate is due to second cousin and other consanguineous marriages .
As a matter of fact , if be the second cousin marriage rate , and the proportion of allogenic element produced by it , and if , be the same functions for third cousin marriages and so on , then equation ( 12 ) becomes and this equation leads to a value of than ( 12 ) .
Thus ( 13 ) determines a smaller value of than ( 12 ) .
Hence if the second and third cousin rates were taken into account the values given in Table II for would to some extent be reduced .
The second and third cousin marriage rates being almost wholly unknown , it has not been feasible to allow for them .
But part of the discrepancy noted between calculation and observation can be explained in this way .
9 .
Finally , it is important to tabulate the relative rates of production of allogenic element by first cousin parents and by the non-related parents , when there is a given frequency of occurrence of the allogenic element in the population at large .
The of first cousin marriage is again taken to have values from 1 to 5 .
The rates of production of the allogenic element by first cousin and nonconsanguineous marriages are respectively and , where is given by equation ( 12 ) .
The relative rate of production is thus simply , where .
This has been calculated for and , and the results are given in Table III .
This table shows while for a pure recessive which occurs in more than 1 per cent. of the general population , a first cousin is at most not more than twice as likely to produce this constituent in its offspring than a anguineous m , in the case of a rare recessive oonstituent with a frequency , say , of 1 in 20,000 , first cousins are from 1911 .
] Stable Simple Mendelian Mr. S. M. Jacob .
lnbreeding in a [ Mar. 18 , 13 to 28 times more likely to have this pure recessive in .
offspring than are non-related couples .
Were Mendelism completely established as a theol.y of reproduction applicable to man , and were it shown that good qualities were absolute dominants and bad qualities always recessive , the above result might legitimately be used to show that first cousin marriages are undesirable .
But this is not so , and , in particular , it eems probable that dominance is not complete , and that there may be some dominant harmful characters and some useful recessive ones .
General Conclusions and Limitations .
10 .
( i ) The not unimportant result has been proved that the more infrequent a pure recessive factor is in any simple stable Mendelian population breeding at random , then this pure recessive is less and less likely to occur in.the offspring of any type of marriage , whether consanguineous or not , though at the same , time it becomes relatively more and more likely to occur in the offspring of related pairs .
In other words , with a less and less frequently occurring pure recessive , the decrease in the appearance of a pure recessive in the offspring of related couples is not so rapid as it is in the case of non-related pairs .
Precisely the same conclusion applies to the production of the pure dominant .
, it has been shown that the relative frequency of the appearance of the allogenic constituent in the offspring of related pairs diminishes by about one-half ' for each grade of cousinship .
Hence , if anything can be first cousin marriages as productive of patent recessive qualities , the argument applies with less and less force to cousin of rades .
( ii ) The general proposition throws a good deal of light on this problem of inbreeding in a simple Mendelian population , namely:\mdash ; The offspring of any system of inbreeding , provided there is no selective marriage or death rate or diffelential fertility , can be expressed in the form That is to say , the offipring of consanguineous marriage can be expressed as a sample of the general population , together with a part due to the inbreeding factor .
To show this , let the offspring of any system of inbreeding consist of a regrouping of the allelomorphs A and , such that the term is added to the general population .
, are functions of 1911 .
] Stable Simple ]Population .
Now , in a population stable both as regards type and number , the total quantity of allelomorphs , whether dominant or recessive , must remain unaltered .
Hence we have at once , and Putting , the result above is obtained at once .
In the case of the children of brother-sister marriages ; for the ildren of first cousins , for , and so on .
thus accentuates both the pure dominant strain and the pure recessive in any stock to the same extent , and at the expense of the hybrid element .
The already noted agreement bveen lihe actually observed somatic correlations and the theoretical Mendelian gametic correlations appears to negative the existence of absolute dominance .
Suppose , then , we are dealing with a harmful recessive characteristic , the result of is that for every individual with patent evil added to the community , useful individual of pure dominant type is also added , and two individuals of neutral type are got rid of .
It is quite conceivable this process might be a gain to the community , owing to a selective action on the patent evil .
It would , of course , in the present state of our knowledge , be impossible to insist on this view : but , at the same time , it must be pointed out that much of the argument against inbreeding fails if the phenomenon of dominance be not complete .
In the particular case in which the has a mean " " utility\ldquo ; between the utilities of the two otes inbreeding is neither vantageous nor the reverse .
( iii ) Further , as Bateson*has pointed out , it may be that some valuable qualities are recessive in character .
In that case , and especially for a very rare recessive , as the analysis shows , inbreeding will serve a useful purpose in the rece , { isive quality to lighG .
( iv ) The analysis has dealt solely with between collaterals of the same rank .
It may , however , be pointed out that where the mal.riages of first cousins once removed are considered , these should probably be regarded in the same as second cousin , so far as resemblance can be used as a test .
For , just as it has been found that first cousins resemble each other to about the same extent as uncle and nephew aunt and nieces , so it appears likely that the correlation between first cousins once removed will be equal to the correlation between second cousins .
But whether this 'Meudel 's Principles of Heredity , ' 1909 .
Dr. J. A. Murray .
Cancerous [ Mar. 22 , equality of correlation exists , and whether the somatic resemblance can be used as a test for predicting the consequences of inbreeding , are points which must be left for future elucidation .
Some of the limitations of the method of this paper have been pointed out , and others are obvious .
extension to the case where more than one pair of allelomorphs is considered might conceivably be valuable when more data have been collected , and should not prove difficult .
In particular , a ethod which assumes the absence of selective mating and ignores the existence of differential fertility can claim no finality .
It is hoped , however , that some of the conclusions are fairly exact deductions from the simple Mendelian theory as it stands at the presen .
time .
In conclusion , I wish to thank Prof. Pearson for the help noted in the paper , and for much stimulating criticism .
Cancerous Ancestry and tloe Incidence of in Jfice .
By J. A. MURRAY , M.D. , B.Sc. , Imperial Cancer Research Fund .
( Communicated by Prof. J. Rose Bradford , See .
R.S. Received March 22 , \mdash ; Read May 4 , 1911 .
) The purpose of these experiments has been the collection of data sufficiently abundant and accurate to determine whether an enhanced liability to cancer is transmitted in the case of mice from parents to offspring .
In a preliminary note in a short account was given of the manner in which these experiments been conducted .
The animals have all been housed and fed in a uniform manner in room .
Tbey have been kept in large cages , which have been cleaned regularly , and the environment has been as uniform as it has been possible to make it .
During the past five years nearly 1600 animals have been bred , the two sexes contribu.ting approximately equal numbers .
Of them , 562 females which have lived for six months or more form the materials of the present paper .
The incidence of the disease is so dependent on the age and sex of the animals that , in order to get comparable groups , only mice of the same sex and of approximately the same age may be reckoned together .
E. .
Bashford and J. A. Murray , " " The Incidence of Cancer of the Mamma in Female Mice of Known Age ' Roy .
Soc. Proc 1909 , , vol. 81 , p. 310 .

