]\gt ; athematical tontrif)to the olntion .
rises in the -percha tubP , the ) on valves inlcreases , the hound b } more marked ; when fluid on the other hand diminishes , ths less distiuct .
If the be placed , the OUllds become vobbly .
The piece of nall d. of a placed over the at ement iuost sitable T bserving solltlds in this ) iment .
The ribed , First suggested by me , VttS subittd , of Dr. Sibbald , to lamcnted fi.iend the lnte ) .
, Sh ) , rr ud to James were satisfied with the ' diminntion of the ullU the sounds closely the ma ] ] , n !
ltl that they bec .
am more or less distinct ortion to of fluid ined i the tnbc .
conclusion , 1 } moved to un 1nnd collby a desire to solution of what to bc olubl e problcm , also by a belief of : of the onnd of vould be of ) the tudy of the ) of diseases this an .
I old further to add the of the ca of .
fi sound of the ] leart ivcu in communication , so that hithcrto accepted , } to oate nlties in the .
of diseases of oser show , howevel , this not to be } , but I will simplify , to .
It fford an of elations of morl)ible , , for a h tolic at the ilsf basc free ; it ' to .
study of th tellsion of to of ilised \ldquo ; LlCL LEE , Bedford S. , of ( April 1 .
ated i the of on the ) COIIIiss A. Lee and Prof. Karl Pearson .
pal'ative correlation in the case of civilised and uncivilised races , and furthel of determining , if possible , any general law connecting relative sexual variation and relative sexual corr lation with the of civilisation , and so with what is probably inversely proportional to the of civilisation , namely , the intensity of natural selection .
The following two pl.incipIes seem to flow from a study of variation in organs of man : ( a ) Civilised man is more uncivilised mall .
( b ) There is a greater equality of variation for sexes in uncivilised than in c.ivilised races .
Civilised woman appears .
on the whole , to be ] ightly more variable than ilised man .
Both these principles are in accordance with the intensity of the for existence\mdash ; and amount , consequently , of natul'al reater for uncivilised than for civilised races , and , further , greater for men than for women in the Iatter races .
The problem of correlation is , however , of a less simple character .
While the action of selection can be shown theoretically to reduce variation , it by no means follows that it l.educes ] ation .
Indeed , selection may crease , decrease , or everse crelation at the very same time as it is reducing variation .
We have tlJen the ving problems guide us in our treatment of actnal statistics:\mdash ; ( a ) Is correlation more intense among ilised than among uncivilised race ( b ) How does the relative COl.relation of the sexes differ in civilised and ulicivilised races ' ( c ) Is there auy prepotency of either sex in the matter of correlatio These the problems which the present calculations designed , to deiinitely solve , but to illustrate .
2 .
Unfortunately , measurements on living members of ilised races are not very for the purposes of colTelation generally very ullterous mea snl'ements of and Qome of bones , but the measurements for each individual are , from the htandpoint , but in numbel Hence , in the case of each determination of a of variation or .
coefficient of COl.relation , the probable el.rol be large .
lt iation in Man , by .
Pearson : " " The Chaltces of vol. 1 , pp. bome cases hnman vlri sexcs dealt elisc of relation of selection to the subjcct of a ) completed .
Actually ] solnc e to higher corrclation among the among the middle pper sscs of civilised Mathematical to the Theory of Evolution .
345 is thus impossible to lay much stress on auy individual relative difference ; we are bound to consider only the general trend of our resuIts .
In order , however , to make the probable er ] as small as possible it is desirab ] in the case of correlation to make the coefficients as large as possible .
Now , for measurements on the skull , the tion is small , but , for the long boils , it is large .
Hence , the lattel will give better results for comI ) arative purposes , if the series be smaller than the former .
They have accordingly been selected for the present investigation .
In all cases standard doviabions .
coefficients of variation , and coefficients of correIation have been ated s by means of the fol.mulae given in a memoir on " " Regression , Heredity , and Panmixia .
The standard deviationh and coefficients of correlation had .
errors detel.mined by formlae of the same ( pp. 266 and 275 ) ; thecoefficients of variation for probable errol , if be the llllmbe of mensurements in the series , and V the cocfficient , 0 .
or sensibly , since vjl00 is of the order , for our pre ' sent investigations .
It will thus be seen that the probable elrol of a coefficient of variation can , like that of a standard deviation , nlount to a considerable of its value if be small .
For xample , if ( case of the Aino women ) it can amount to about 10 per cent. ; while for ) ( case of the French men or lvomen ) it will amount to .
cent. The probnble errors of the standard deviations ] also in the two cases amonnt to 10 and per cent. espectively .
On the other , if be as small as ) , , the probable of the coefficent of elation will only be per cent. of its value if be as lal.ge as , and only per cent. of its value if .
On the othe ] .
hand for the probable rors of for the cases of and are respectively and ) per cent. It thus appears that if be large , or even fairly laroe , it can be ( Ietermined with far less probable va ] lation .
The results for correlation are thus likely in the esent case to be significant than tlJose fiation .
3 .
As to the itself French measurements from a work by ] .
Rollet : 'De la lIenburation des des Iembres , ' published in the 'Bibliotheque ] Lyonb , The bodieH from the taken having paf ; sed through the post-mortem , the sex known without question .
The member } ) * ' Phil. 1 ralls pp. } VOL. Miss A. Lee and Prof. Karl Pearson .
except in the few cases its length was not recorded .
In these cases in orde1 not to diminish the already small series , the left member was taken .
Of the two measurements of the elnur given by Rollet the longer was taken .
The measurements of the Aino bones were taken from a work by Koganei iu the ' Mittheilungen aus der Medicinischen Facultat der K. Japanischen Universitat , ' Bd .
2\mdash ; Tables , Tokio , 1894 ; and the same rule observed .
The Japanese returns , while more extended in the variety of bones measured than the French , yet contain fewer illdividuals , and in many lespects do not seem to us so relia , ble as the latter .
* In both cases the maximum length of femur , humerus , radius , ulna , fibula , and clavicle , Fib .
, and Cl. ) were measured .
For the Aino we have in addition to the longhs ( and ) of the radius and tibia from centre to the length of head and neck of femur , and the scapular index ( Sc. ) or ratio of breadth to length of scapula .
The Aino measurements give the maximum length of the tibia as wsll as the from centre to centre .
The French give the of the tibia , excluding the spine and including the malleolus ( T ) .
In order our results might be of service to Mr. E. Warren , who is dealing length with the Libyan long bones , the radius elations for the Aino were calculated for , and not for , as they are in the case of the French .
The results for French and Aino are not stlicbly comparable in this case .
The same holds for the measurement of the tibia .
Probably Rollet 's COl'responds more closely to Koganei 's measurement No. 13 than to his No. 14 .
Both and are is less fibula in length .
We have dealt with the Aino , howevel\ldquo ; for the purpose of comparison with the Libyans .
The following ables give ( .hief esults :\mdash ; * Onc humerus noticed to be very it be excluded as " " discordant observation\ldquo ; by it it was tlftel ' found to be described in the text as .
Tlle merus ( No. 31 ) bhould be 279 and not 379 cm .
long .
The fcmale ( 270 cm .
) has been rejccled .
It is out of all to the ] ) , as judged by the hole range of ' ( .
yet standing its ] : jbes it ( p. 127 ) as much bent .
The mnle fib , No. 19 , is Withont it the fib tion llathematical to the Th of Evolution .
347 atical Conlributions to the Theory of olutio .
349 Miss A. Lee and Prof. Karl Pearson .
Table .\mdash ; Sexual Ratios : 8 to S. Miss A. Lee and Prof. Karl Pearson .
Table Sexual Ratios : 8 to 9 .
relations .
Mathematical tributions to the ovolution .
Table \mdash ; Racial Ratios : French to Aino .
Correlations .
Rate of progression- of females ovel males in tion the tables illustrate at once how little can be judged the of variation from any single measul'ement when the serieb are so small , and the measurements do not entirely respond .
We forced to arguc only from the bo.elleral trend of the whole .
On the other hRnd , the exccption of cla ] and scapular } , the errors iu the ction { sufficiently for inferencos to be We will deal with the general to } , and then with those bearing on race .
6 .
Sexual 11 ' .\mdash ; In the sexual ratio for the mean of bone ter nnity . .
both rench and Aino together , we have series iving nlcan sexual the bones in civilised and uncivilised man of .
This well the value mined for a serics of rent in the paper on ' tion in } ' p. 374 ) eferred to above .
Turnlng to xual lriation , find , for rench and ether , .
The *So far as these .
to the natnre of relal i betw long bones , we have placed our the of } rell , ho dealing with ] the long bones of the L ) Miss A. Lee and Prof. Pearson .
detel.minedfi.om about 155 series ( loc. cif .
, p. 373 ) was .
A study of the bones thus seems to confirm the result previously eached , that woman is , on the whole , slightly more variable th man .
Lastly , turnin .
the sexual ratio in correlation , we find that in 14 out of 22 cases , ) is more highly ted than man , the mean sexual rabio in correlation So far : then , as the present results go , they seem to indicate that man , while larger , is less variable , less correlated than woman .
When we take , howcver , into consideration , there are additional factors which somewhat modify these conclusions .
In the firsb place there is , of course , no suggestion that the modern ]rench lineal descendants of Aio , but it does not seem improbable position that the French have at one time ssed thro a somewhat similal to that condition in which llow find the Aino .
Assuming this the purpose , at rate , of comparison , we find that transition the to the civilised condition is accompallied by -mnrked rhanges in the sexual atios .
They as follows:\mdash ; to in size woman .
Woman tends to in variability on ma ] ] .
( iii .
) Woman tends to gain in correlation on man .
In .
words , primitive man and woman are more equal in size , variability , and correlation than highly civilised man and woman .
Relative to woman , man 's in size has been anied by aloss in variability and in the correlation of his parts .
So .
as size and bility .
are concerned , results in accol.dance with the .
vider series of mcasurements dealt with in the paper on ' Variation in Man and Woman .
' The result tion was there ested on very evidence .
Before we 111 any gestion to bhe ring of results om uatu al selection it will be well to mine the racial ratios .
7 .
if ?
acial \mdash ; In means both ( 3 and ench a in measurement larger than ?
and 2 Aino , but the for males than females .
ioxlS tho 8 ?
cnch lIave 1 ) gressed iu case on ( ; and 9 Aino ( the tibia i11 the male and in the lc ftre ) rently exceptions ) .
But while felnales have progressed less ] in absolute , they Last elation the rench males more the Aino males in orgnn , and less in six mean being almost eqnality of tion .
othc rench females on infel.ior to Aino in olte hnnl sensible in Contributions to olution .
355 conclusion would then be that , irlcreased civilisation , size*and val.iation tend to increase ; while correlation , to judge by the males , is ; to judge by the females , tends to It will be found somewhat difficult to reconcile these resuIts with simple tions of the pl.inciple of natural selection .
In the first place in } , variation ulldoubtedty suggests alessen]ng of the stru .
cxisbeuce , and there can be no question that this increase has .
on amoIlg civilised races ( see tion lIan and l of the has probably been reater f man ; hence the principle of natural selection } ) th } of lised .
The increase in with civilisation seems , on the average .
able .
But is it the effect of lessening the for cxiqtenc sibilities m be summed up as follows:\mdash ; The civilised races may have survived owing to their sizc .
It may be result of the ] in the past .
To ) must be objected that the increase of ppears to be a progressive ge still going on , and yet ease of variation should show a strnggle for existence .
( b ) effect of suspending latural selection may be to inclease size .
This would be a blow for panmixia , for we have expected a to the smallness of the more imitive races .
old unexplained the small of vomen ontpared with ) men , for in .
case we argue the Vfll.iation tbab the glc for existence is tively less than the cabG of ) } ( c ) si of the ilised races bc to better food lpply and bette ] .
) .
: in sborb it be due , not to , but to ) ditiollS of ypothesis does volv( : the asRllml)tion that arc inherited .
Dimillisl ] ) food hupl ) and abolish physical aining and the size vould to leve ] } which selection had ieft it .
Physical in civilised j being llsually in tlIe case of man than volnan would , pel.haps , why man has ) essed more ) in size seems possible , tion as a measure of ensity of ) , to the relative eases i of man with any dil.ect effect of ural stion .
is only/ ( cn ( ' , not in id fen is .
lban of the } , and of itants of Canary other hand , ) ) to be ] lcs the Libyan , { the merus is some omen appear in all long less than the L ) Mathematical Con tributions to the of tion .
8 .
To sum up , then , the following esults seem suggested by these measurements.u .
Civilised man has progressed generally on primitive man in , variation , and correlation .
( ii .
) This prog ] ession can hardly be accounted for } ) increased selection ( because of the increased variation ) , nor by decreased selection it is inconsistent with the relative changes iu male and female size ) .
It might possibly be accounted for by decreased selection and improved physical conditions .
( iii .
) Woman is more variable than man in civilised races .
Woman is more ated than man in civilised races .
In uncivilised races the sexes are more nearly equal in the mattel of size , variation , and ] ation than in the case of civilised races .
It is impossible to say that civilised woman is nearer to the type than civilised man , for while civilised man differs more from the primitive type than ciilised woman , so far , plobably , as absolute size is concerned , he has made only about half her gress in variation , and hardly any at all in corre(vii .
) The causes ( e.g. , lessening of which tend to increase variabion may also crease correlation .
In other words , the intensity of the for existence is not a measul'e of the sity The mastll.ements made Wal.ren on th yans , results of which he has kindly , arc , on the whole .
in accordance with the above conclnsions .
He for the Mean of the sexual ratio of the nlcans the val.iations .
the COl.relations The correspollding quantities for the French we concluded , that in uncivilised to civilised peoples , from Libyan to rench , the men gain on omen i size\mdash ; very slightly , and the women gain upon the men nlfilkedly in tion and relation .
These Cbults are lllerely gestionb , but bhey OSj to emphasise the nnce of a careful bones of , say , 100 mcmbers of th sexes of civilised civilised races .
the former t be l the lleed of ) on , in { , sa } { , may ally tcncl 10 red } tion lv other } ) humcrns the of ontagin n of surcments similar to those of 11 .
Rollet for both nglish and .
The value of such statistics for comparative plrposeb -ould be great . .
' On the Nature of the Contagium of Rinderpest .
iuary mmunication.\ldquo ; By ALEXANDER GTON .
F.R.S.E. , Director Colonial Institute , Cape Colony .
icated by Sir JANES , LL. D. , F.R. farch 2 June 3 , 1897 .
In the following pages it is proposed to communicate to the Royal Society the of experiments made in Soubh infectivity of the blood of animals affected with Rindel.pest .
experiments were all made on cattle kept undelonditious in ich accidental spoutaneous infection could with be excluded .
These experiments had been collcluded in 1896 , before the arri .
R. Koch in South , and their ults had been cated to on his ival .
1 .
The blood of an animal ill with rinderpest , when taken febrile stage or ious to death , and injected subcutaneously intravenously into cattle , produces the disease\mdash ; rinderpest , provided the ) is prevented coagulating .
2 .
The onset of coagulation and actual coagulation of the blood exert a marked dtuctive infloence on the ulence of such bloed .
3 .
The best method of obtaining virulent blood is to it tically from the vein of an ill with rindel.pest , to mix it im1nediatcly a 1 per cent. solution of citrate of ) ] the previously well ilised , in the proportion of arts blood to 1 part of citrate of potash solution .
Such blood , been flnid .
4 .
citrate of of blood proves virnlcnt first few days , generally not six days ; ftel .
six days ' keeping ]cnce becomcs rapidly eakened , so that days the blood mixturc is inert .
5 .
Admixtnrc of to citrate blood does not ) interfere with the virulence of such blood .
to blood does interfere tb ulen latter itccount of the coagulation of ) 6 .
mucus of an infected when used and ubbed into the nostl.ils of cattle , produced in all instances typical rindcrpest .
We have nevel had a failuro in ) to produce the disease by this mcans .
By he mucus , for a few ] ) ours , its irulence become markedly less .

