INHERITANCE IN MANKIND 51 



special cases ; but to measure quantitatively the heredi- 

 tary influence, it is necessary to deal statistically with 

 large numbers of such pedigrees. Moreover, it is not 

 enough to study diseased stocks only. They must 

 be taken as parts of the general population, and our 

 statistical studies must deal with fair samples of the 

 population at large. 



The possible pitfalls of too restricted a view are 

 better illustrated by the more difficult case of tuber- 

 culosis, which we have dealt with last, than by the 

 more rare condition of deaf-mutism, where no question 

 of infection can arise. 



Tuberculosis is very prevalent. It affects fatally 

 some ten in a hundred of the population. Hence in a 

 sample of the people taken at random one or two in 

 seventeen might be expected to suffer from the disease. 



Although these numbers would be exceeded in 

 chance cases, the probability of as many as fifteen out 

 of seventeen in a group suffering would be exceedingly 

 small. Nevertheless, such a conjunction would be 

 possible, and a few such cases taken alone could not 

 settle the question. 



Again, since tuberculosis is infectious, it could . be 

 argued that members of the same family might more 

 readily infect each other than those outside. It should, 

 however, be noted that, as tuberculosis is, in general, a 

 disease which develops in early mature life, and affects 

 comparatively few children or old people, it is less likely 

 to run through a family by infection than some other 

 complaints. It would be much more likely to be passed 

 by infection between husbands and wives. Some in- 

 fection in these cases probably does occur, but a 



