XXI .
" Supplement to Two Papers published in the Philosophical Transactions ( 1820 and 1825 ) on the Science connected with Human Mortality .
" By BENJAMIN GOMPERTZ , Esq. , F.R.S. Received June 19 , 1861 .
( Abstract .
) The first of the two Papers referred to is entitled " A Sketch of an Analysis and Notation applicable to the estimation of the value of Life Contingencies .
" It is especially devoted to the explanation of a new notation and a new mode of analysis applicable to the subject .
The second Paper , entitled " On the Nature of the Function expressive of the Law of Human Mortality , and on a new mode of determining Life Contingencies , " treats , in its first part , of the equation L , =d .
g q , in which L , denotes the number of persons who would be living at the age x , out of the number of persons who may have been living at some given common previous age , in which d , g. , q may be considered constant quantities through a long series of years , and the equations to give the real rate of mortality during that period within a very satisfactory approach to coincidence , say , for instance , from the age ten to the age sixty ; a theorem from whence the number living at an assigned age may be derived , by what the author calls the vital rule of three , from the number living , for instance , at the age twenty , forty , and sixty , given in a table of mortality .
But the author observes that though the above formula with suitably determined constant values of d , g , q appears within a great approach of the observed result , to agree with the tabular data through so long a period , still that actually d , g , q are not constant quantities , as every different selection from the table of the living to which the formula is applied , for the purpose of determining the three constants contained in it , even if that table throughout were an accurate representation of the law of mortality , which no table based on obtainable data can be expected to be , would give different values of d , g , q , as they are not really absolutely constant , but slightly variable throughout , because the formula L.=d.g q does not represent perfectly the law of mortality , if d , g , q are perfectly constant ; and the author complains of its misinterpretation in this respect by a subsequent writer , who lays claim to discovery on this subject .
The other part of the Paper of 1825 is devoted to a mode with accompanying tables , to enable a calculator with any given table of mortality to calculate the value of an annuity on any number of joint lives to within any degree of accuracy at pleasure ; provided the periods , if long , are divided into parts , The present supplement , written so many years after the original papers were communicated to the Society , appeared to the author to be called for in order that he might further illustrate the subject , and add his later speculations and improvements .
The supplement the author commenced writing about two years ago from collections of his manuscripts , but was prevented by continued ill-health from proceeding with it ; but having received an invitation last year from the International Congress to assist in their meritorious labours , and being unable from indisposition to attend the Meeting , as he stated to the Congress , he offered to send them some hints respecting his recent labours , the results of which he intended to present to the Royal Society , if his health should permit him to finish the papers .
The hints which he gave , which were honoured by a place in the Reports of the Congress , were , he believes , deemed interesting , but he thinks they were sufficiently separated from the strictly mathematical part of the subject , and also from the most important portions of the result , to allow him to consider the present paper a new work , or sufficiently new , to be thought worthy of presentation to the Royal Society , and especially because since the notice above alluded to was furnished , the author has been able to introduce important improvements .
The original formulae give one uniform law of mortality from birth to the utmost limit of the table ; but in the case of the table presented to the Congress , and of the present further improved formulae , if a comparison be made , for instance , with the Carlisle table from birth to the age of 100 , and even from birth to one month , to two months , to three months , to six months , and to one year , where the tables appear to be so irregular , and formed on no law , the result appears to show that a law really exists and is available .
The author shows the vast use and applicability to solutions of all intricate inquiries of life contingencies of his present deductions .
He gives a theory , which he believes to be quite new , which he calls special , single , and specially influenced contingencies ; a subject which he states he had not lost sight of when officially engaged in the science of assurance , as he considered that all cases of connected lives , where there might be a connexion of any sort , of influences acting with respect to the probability of co-existence , all questions solved without reference to such influence , were faulty , and might be importantly faulty .
He states a striking instance with respect to this observation as follows:-If it be required to assure a sum on the joint lives of A and B , that is , to pay the sum on the first of the death of A and B , whichever may be that first , and if in another case it were required to assure the sum on the death of B in particular , provided he should die in the lifetime of A , it is evident that if A and B are of the same age , and not subject to any influential connexion , that if they are both subject to the same uninfluenced mortality , the price of the first assurance would be exactly double the other .
But if the assurance were for a very short time , for instance , in the time a ship was passing from Dover to Calais , and they were both going in the same ship together , the risk of the ship being wrecked would affect them both equally , and the chance of their separate escape would depend on circumstances , for instance , one being a better swimmer than the other ; but there would be a chance that neither of them could escape , and as far as that chance is concerned , the assurance on one of the two , whichever it might be , who died first , the assurance on one in particular , or the assurance on the longest of their lives during that voyage would be all of equal value ; and the method the author has adopted for all uninfluenced contingencies would be available in all cases of influenced , by the introduction of certain formulae which the author regards as of a very interesting nature .
In addition to the above particulars , the author presents a theory of sickness , which appears to him to be a near approach to actual statements of sickness occurring among Friendly Societies .

