4:)S4 Assovtatire 2Iutin(j in Man 



(Irawn from thesc, and the size of inost such churchyaids is suffioiently iiidicated 

 by the fact that 37 uf thutn yielded ouly S9U legiblc iuscriptiuns of the kind 

 required. The only towii <it' any size included is Bicester; in this place the 

 chuirhyard is sharply dividi-d iuto an older portion, containing graves niade before 

 18.')0, and a newer portion; the older portion only, as nnlikely to be influenced by 

 modern nrban conditions, was tabulated. 



The niass of the records relates undoiibtedly to the farniing popnlation, iiicliuling 

 a considerable proportion of the labouring class, who obtain stone tombstones niore 

 easily in (.)xfordshire, where building stone is cheap and easily cut, than in many 

 places. The proportion of resident gcntry, of clergy, and of professional men, 

 is greater than in the Yorkshire records. The data were coUected by F. Biichanan 

 and W. F. R. Weldon, tabulated and roduccd by K. Pcarson, E. H. J. Schnster, and 

 W. F. R. Weldon. 



(c) Records of the Society of Friend^s. Collected by Mary Beeton froni 

 pedigrees, chiefly printed books at Devunshire Honse*. The material was 

 tabulated and rcduccd by M. Beeton herseif. Whilc locally far niore scattered 

 than tlic material considered in (a) and (b), the Society of Friends fornis a 

 remarkably homogcneous body. Special features differentiating these data from 

 the above lural district data are the far lower averagc age at death of both 

 hnsband and wife, and, what does not appear in the results connnunic;ited in 

 this paper, the occurrence wiih considerable frequency of remarkably early 

 marriages. 



(d) London Cemeteries. Data collectiHl by Mary Beeton from St Mary-le-bone 

 Cemetcry, Highgate Cemctery, Hanipstead Ceinetery, aud St John's Graveyard, 

 Hampstead. The tabulation and reduction are due to M. Beeton. 



(4) The abovc data cover a long ränge of yeare, both for husbands and wives, 

 the lowest age at death being 18 ycars, the highest 107 _vears; these ages are 

 subject to some uncertaint}-, because a man recorded as dying " in the Oüth year 

 of his ;ige " may have died at any tiine betvveen his 89th and 90th birthdays, and 

 a man dying " aged 90" may have died at any time between his itOth and his 9Ist 

 birthdays; fnrther, it is by no means certain that the distinction between these 

 two forms of statement is rigidly observed. In onler to employ a unit of age 

 which should not be affected by this nncertainty, and at the same time to Jessen 

 the work of tabulation, a tive year pcriod was adopted as the unit of age, both for 

 husbands and for wives. The first question which arises is the effect, which may 

 have been produced on the data, by the use of so large a uiiii : this was tested by 

 F. E. Lutz, who determincd the mean age at death and the Standard deviation for 

 husbands and wives from the Wensloydale district and the correlation between 

 them, using first five years and tiieii ten years as the unit age; using Sheppard's 

 corrections for the moment forraulaef he obtained the results given below; these 



* See Jt. S. Proc. Vol. 07, p. IGO. We dcsire again to express our thanks to Mr Isaac Sharpe for 

 is courtcona assistancc. 



t See Biometrika, Vol. i. p. 273. 



