Danforth: Resemblance and Difference 



405 



similarity in some respects and dissimi- 

 larity in others, for it may be assumed 

 that the traits are not all quite identical. 

 Stated in other words, there would 

 seem to be evidence that all bodily and 

 mental traits tend to be similar in cer- 

 tain (mostly uniovular) twins, but that 

 it is not unusual in twins of either type 

 to find dififerences, now in one trait and 

 now in another. 



EXTENT OF SIMILARITY IN 

 BIOVULAR TWINS 



In regard to the second question, it 

 may be assumed that biovular twins will 

 on the whole resemble each other to a 

 degree about equal to the average for 

 all children of the same family. In indi- 

 vidual sets the degree of resemblance 

 will vary widely, since the possible com- 

 bination of traits from the maternal 

 and paternal germ plasms are very 

 many.^ It is conceivable that occa- 

 sionally (with our population and birth 

 rate, perhaps once in eight or ten years) 

 a pair of biovular twins would be born 

 with identical germ plasms. More fre- 

 quent should be biovular twins differ- 

 ing in only one chromosome, and pro- 

 gressively more frequent the pairs dif- 

 fering by larger and larger degrees until 

 the average of fraternal resemblance is 

 reached. Then progressively less fre- 

 quent should be pairs differing to in- 

 creasing degree below the fraternal 

 average. 



Since most pairs of parents have 

 many points in common, perhaps select- 

 ing each other on the basis of com- 

 munity of physical as well as mental 

 traits,^ the possible average of com- 

 bination of traits within a given family 

 is somewhat reduced and the likelihood 

 of children with similar germ plasm 

 correspondingly increased. If two very 

 similar children are born a few years 

 apart, their likeness may be easily over- 

 looked—the photograph of a boy at 

 eight and the same boy at fourteen may 



show no more than a family resemblance 

 — but when such similar children hap- 

 pen to be twins of the same sex, even 

 though the degree of resemblance is 

 such as could easily be matched in other 

 brothers and sisters who are not twins, 

 the similarity of age and frequently of 

 dress immediately becomes noticeable. 

 Such twins will pass among their 

 friends as strikingly similar. It is in- 

 teresting to note that in submitting 

 photographs, many twins explain that 

 they are really more alike than their 

 pictures indicate. On a priori grounds, 

 then, many pairs of biovular twins 

 should be expected to show resem- 

 blances, either real or apparent, that 

 would be much above the average for 

 the ordinary run of brothers and 

 sisters. 



TWINS OF DIFFERENT TRAITS OFTEN 

 OVERLOOKED 



It will no doubt be objected that, for 

 every pair of biovular twins which 

 show a resemblance greater than the 

 average fraternal resemblance, the laws 

 of chance lead one to expect that there 

 will be another pair that will show a 

 corresponding dissimilarity, bringing the 

 average resemblance down to that for 

 brothers and sisters in general. Assum- 

 ing that the distribution of traits in 

 biovular twins at birth is in accordance 

 with the laws of chance, and ignoring; 

 the possibility of subsequent environ- 

 mental selection (e. g., by disease) 

 against dissimilarity in children, there 

 remains an important factor to be con- 

 sidered in any study of twins that have 

 passed the age of infancy. This is the 

 likelihood of overlooking twins who do 

 not closely resemble each other. Chil- 

 dren who differ widely in general ap- 

 pearance, size, mental capacity and 

 tastes are less likely to be together, 

 more likely to leave school at different 

 times and, curiously enough, often take 

 little interest in the fact that they are 



5 On ilie assumption that chromosomes maintain their individuality and that "crossing 

 over" docs not occur, there would still be no less than 4,096 possible kinds of germ cells 

 to be produced by a given individual of civher sex. 



6 H?.Tis, T. Arthur: "Assortative Mating in Man." Pap. Sci. Monthly, vol. Ixxx, 

 pp. 476-492, 1912. 



