198 G. Udny Yule. 
It is not essential for the application of such ‘‘estimating 
equations ” as ( 1 ) or ( 2 ) to the study of heredity that the variation 
of the character concerned should be strictly continuous as in the 
case of stature in man. The method is equally applicable to any 
form of variation in a scalar series, even when such variation 
proceeds by a series of discrete steps. The variation of petals, 
sepals, and other floral parts, and of numbers of offspring, may he 
quoted as familiar instances. Even in such cases as stature, dis¬ 
continuity is, in point of fact, nearly always introduced by the 
observations being grouped, or the measurements being only taken 
to some considerable unit of the scale, e.g. to the nearest half¬ 
inch or quarter-inch. Between real continuity—a continuity that 
appears to be unbroken with the most careful measurements 
possible—and the discontinuity of a scalar series proceeding by 
successive equal units there is therefore no important distinction . 
The same method and conceptions apply, the same law of regression 
must hold. 
A distinction arises, however, if no scalar series exists, but the 
race is simply divided into two exclusive classes, the one possessing 
some attribute, the other not; as one may divide a race of men into 
deaf-mutes and normals, sane and insane. In such a case the 
statistician again speaks of the attribute as being inherited , if the 
character of the parent enables one to estimate the character of the 
offspring more accurately than would be possible from a mere 
knowledge of the general characters of the race. If the two classes 
be termed A’s and a’ s, then the attribute is inherited if the per- 
of A ’s amongst the offspring of A’s is larger than the percentage 
amongst the offspring of a’s. This is, as before, individual heredity. 
When the biologist speaks of the transmission of an attribute 
common to all the members of a race as heredity (as the flowers of 
one race of plants may be white, those of another pink; the stems 
of one glabrous, those of another hairy), he is dealing with a 
quite distinct aspect of the phenonemon. I do not, of course, 
object to such an accepted use of the term, but wish to emphasise 
the distinction, 
Whether, in fact, we deal with continuous variables or attri¬ 
butes, all the individuals observed, in any one case, must be 
members of one race, or else the two phenomena of race-heredity 
and individual heredity are superposed and confused. If, for 
instance, stature-measurements on a tall race and a short race were 
mixed, it is conceivable that there should be no individual heredity 
