452 REPORTS ON THE STATE OF SCIENCE. 



information as may enable the Committee to recognise the disease on 

 which he is reporting. 



(6) Disease statistics of population. 



Statistics concerning the prevalence of or the mortality from disease 

 are of little value by themselves, nor can their variations be accurately 

 compared with climatic changes, or, indeed, with any other variable, till 

 their primary relationship to the population from which they are drawn 

 is established. Further, the incidence of a mortality is often selective in 

 the same race, determined usually by age, often by occupation, sometimes 

 by sex. 



The requirements for the satisfactory comparison of disease data may 

 be set forth in tabular form : — 



Total cases of disease \ Reniiire /Total population from which they 



Total deaths from disease J ^ ' ' I. were drawn. 



Cases by ages "[ i Population by ages, and possibly 



Deaths „ J ' ' • " • "I total deaths by ages. 



g-^y-} ....... Population by sex. 



Where the disease appears to be limited, chiefly or entirely, to one 

 section of the population, these statistical details should be given for the 

 sectional as well as for the total population. 



Racial differences in incidence and mortality are most important, and 

 can only be elucidated by comparative statements in which all the par- 

 ticulars enumerated above are comprised. 



It may sometimes be impossible to obtain exact statements as to age 

 aistribution of the population, but the best available information should 

 be given. There may be in certain localities a temporary change in the 

 age distribution of the population, where for religious, economic, or 

 climatic reasons, the tide of immigration (usually mainly of young adults) 

 flows at regular or irregular intervals. Probably this in most cases affects 

 the annual rather than the monthly ratios, but it is a point which is 

 worthy of consideration. 



(c) Disease statistics. 



These fall into two classes : — 



(i) Where cases and deaths are recorded. 



These include statistics from hospitals (civil, naval, and military), and 

 from prisons and gaols. The published data from these sources usually 

 give annual values for all and monthly for certain diseases. They 

 include, of course, all cases and deaths which come under observation ; 

 they are complete as far as they go. But the relation of these hospital 

 cases to the population varies. In naval and military statistics the 

 whole sickness of the navy or army is dealt with, hence the actual inci- 

 dence rate on the population is known. The deaths, however, do not 

 represent the total mortality due to sickness contracted in the service, as 

 a certain small proportion of men die after invaliding out of the service 

 from disease actually contracted in the service. 



Civil hospitals deal only with a part of the sickness of the population, 

 and the relation of this part to the total sickness of the population is a 

 very variable quantity. Hence, normally, the number of attacks in the 

 civil population is not known, The deaths in hospital also do not 



