152 THE BIOLOGY OF TWINS 



pure for this character. Otherwise why does a mother 

 with doubhng always have offspring that show some 

 form of doubhng ? 



Parallel somatic and germinal segregation can be 

 explained only on the assumption that the segregation 

 mechanism operates at a period prior to the differentia- 

 tion of germinal and somatic cells, and this must be 

 during the early cleavage stages. 



Mirror-imaging and symmetry reversals as the result 

 of polyemhryony. — So frequent is the occurrence of 

 mirror-imaging among armadillo quadruplets that it 

 must be conceived as causally related with polyembry- 

 onic development. The two phenomena are so closely 

 related that it is my belief that the occurrence of sym- 

 metry reversal or mirror-imaging in twins or double 

 monsters may safely be taken as a criterion of their 

 monozygotic origin. Only, however, when there appears 

 some asymmetric feature like unilateral doubling is 

 mirror-imaging recognizable. If both sides of twin 

 individuals are exact bilateral duplicates no symmetry 

 reversals would be possible; hence we must depend upon 

 the unilateral appearance of such features as doubling 

 for signs of reversed symmetry relations. 



All grades of mirror-imaging are found in armadillo 

 quadruplets. There may be mirror-imaging between 

 individuals of opposite pairs, but this is much less 

 common than imaging between twin partners derived 

 from one half of the Qgg. 



A still more frequent type of mirror-imaging is seen 

 between the antimeric halves of a single individual. 

 These facts lead to the conclusion that the symmetry 

 relations among quadruplets are the result of an intri- 



