VARIATION AND HEREDITY IN TWINS 131 



0.5638=1=0.0597. Making allowance for i)ro]Kibility 

 of error, the coefficients of both sexes are practically 

 identical; the coefficient is about 0.5, which is ]\\>i 

 what we should expect if mother and father contributed 

 equally to the inheritance of these characters. Pre- 

 sumably, then, the coefficient for fathers and offspring 

 would also be o. 5 or thereabout. It will be noted also 

 that the males inherit from mothers just as strongly 

 as do the females, which goes to show that we can ignore 

 sex in dealing with the inheritance of scute characters. 

 Practically the same degree of correlation exists for 

 the number of scutes in the other four parts of the 

 armor. All of this evidence simply means that, on the 

 average, embryos show as much paternal dominance 

 in scute numbers as they do maternal. Remembering, 

 then, that the degree of resemblance between mothers 

 and their polyembryonic sets of offspring is about 0.5, 

 it will be interesting to compare w^ith this the degree of 

 resemblance among the quadruplets themselves. 



Correlations among individuals of quadruplet sets as to 

 numbers of scutes.- — ^Using the same statistical methods as 

 in determining the heredity between mother and offspring, 

 we find that for total numbers of scutes in the banded 

 region there is a coefficient of correlation for 56 sets of 

 male quadruplets of 0.9294=1=0.0057, and for 59 sets 

 of female quadruplets, 0.9129=1=0.0059. This degree 

 of correlation is extremely high and it has no paral- 

 lel among interindividual correlation coefficients. In 

 fact, the members of quadruplet sets are as strikingly 

 alike (as closely correlated) as are the paired organs of 

 single individuals or as the right half of a single 

 individual is like the left. The correlation constants 



