148 H. H. NEWMAN 



Perhaps the most striking revelations had to do with the occur- 

 rence of rare atypical arrangements of bands. In some cases a 

 certain peculiar band arrangement was repeated with remarkable 

 fidelity of detail in two, three or even four individuals of a set, a 

 circumstance seeming to indicate unequivocally that the character 

 in question must have been in some way predetermined before 

 the separation of the embryonic rudiments of the four members 

 of the set. Equally striking and perhaps even more puzzling 

 were the cases in which certain single scutes, peculiar in some re- 

 spect, were seen to recur in two or more fetuses in almost precisely 

 the same location. Such units of organization were referred to as 

 probably the smallest characters capable of hereditary control 

 and transmission since thej^ are of the same order as hair groups 

 in other mammals. 



Throughout the paper attention was repeatedly called to the 

 fact that the fetuses are arranged in pairs, one pair attached to 

 the right hand placental disc and the other to the left, and that 

 the resemblances were closer between the individuals of a pair 

 than between those of opposite pairs. The pairing was explained 

 by assuming that each pair came from one blastomere of the two- 

 cell stage of cleavage, an assumption which now seems unwar- 

 ranted and unnecessary. 



The paper summarized above called forth comment both com- 

 mendatory and critical. By some the work was considered highly 

 suggestive and worth}^ further analj^sis, by others it was rele- 

 gated to the category of papers that missed the mark and had no 

 especial significance. Among the criticisms and helpful sugges- 

 tions that have come to me from various sources are the following: 



1. A well known biometrical expert writes frankly that he is 

 unable to place any reliance on the accuracy of coefficients of 

 correlation derived from only twenty sets of quadruplets, espe- 

 cially since the probable error of the determination has not been 

 worked out. This objection should be satisfactorily met in the 

 present studies, for nearly seven times the original number of 

 sets of quadruplets are now available. Curiously enough the 

 results of determinations of constants, when this larger number was 

 used, do not differ materially from those obtained from the smaller 



