878 H. H. NEWMAN AND J. THOMAS PATTERSON 
single individual and that we are dealing mae a special case of 
intra-individual variation. 
Were the phenomenon of specific polyembryony in need of 
any further support, the discovery that the degree of correlation 
among the foetuses of a set is the equivalent of that found to exist 
between the right and left sides of the same individual, would 
in itself constitute a demonstration of its validity. 
TABLE 7 
Showing the variability of seven pairs of duplicate twins in terms of the coefficient of 
MENS INS eran 
{ 
DESIGNATION OF TWINS | COEFFICIENT OF VARIATION 
WGISTIT aI Secs ah Oe eee eae UR RE iy aR ara eet A eee ae | ike 
WEEMOM BTCA Pea ye eke tee dee ert en ae ne ene erento | 0.22 
MELMONSS (HID Hirk Lette heel bes Rrra ae eee es Cree | 0.34 
Wal ier Sa Ret ih taeda ci eons earn, es a Re ne 1.438 
Nall Ve Nes ech ie epee Rp Feet ems Mar vs eee Rm aA 1.65 
VSG OTS SHI Ties, c Meee ee cee se PPS cee oo ey, Nee irae, 
Wiltlenesriny ic bcc alt Ars cit oP eee, OF Cano) REED Soden Ls iley 
IAW ETT earn Mee Sea ce ees nn Mrs cube aes Ba ae ee cms ee | 1.06 
In order to be able to institute a comparison between the 
strengths of hereditary control exhibited by armadillo quadrup- 
lets and human duplicate twins it becomes necessary to employ 
some method which does, not involve the use of the coefficient 
of correlation, for the twin data is too scattering and diversified 
for our purposes. The usual method employed by the various 
writers for indicating the degree of resemblance between twins is 
that mentioned in the introduction, namely the per cent differ- 
ence method, but this can be used only in the comparison of pairs. 
A better method for our purposes involves the determination of 
the coefficient of variation for each pair or set and an averaging 
of these constants. Table 7 gives the data on the seven pairs of 
human identical twins referred to above. In table 8 are recorded 
the coefficients of variation determined for the twenty sets of 
foetuses. It will be noted that the degree of resemblance among 
the quadruplets is, on the average, nearly twice as close as that 
of thetwins. Thisis perhaps no more than we should expect, even 
