822 



SCIENCE 



[N. S. Vol. XXXVII. No. 961 



only be explained by an exact knowledge 

 of the grandparents. We have already 

 seen a concrete example of this ; but before 

 returning to it the essential result may be 

 explained by means of a diagram. Let us 

 consider the case of a hybrid that contains 

 four characters, which we will designate as 

 A, a, B and b. The hybrid AaBb can be 

 made in two ways, according to the compo- 

 sition of the parents. First, one parent may 

 contribute AB and the other ab ; or, sec- 

 ondly, one parent may contribute Ab and 

 the other aB. These two crosses seem to 

 give precisely the same result, AaBb. The 

 hybrids produced by the two methods look 

 exactly alike; they produce the same kind 

 of offspring (grandchildren of the original 

 forms). The remarkable fact is, however, 

 that in some cases (probably in many) the 

 offspring of the two kinds of hybrids differ 

 in respect to the numerical proportions in 

 which different combinations of the grand- 

 parental characters appear. In both cases 

 the grandchildren are of four visibly dif- 

 ferent types, AB, aB, Ab and ab. Follow- 

 ing the first cross, however (AB X ab),the 

 classes AB and ab are in great excess 

 among the grandchildren, sometimes in 

 very great excess; while following the sec- 

 ond cross (Ab X aB) it is the classes Ab and 

 aB that are in excess. In other words, in 

 each case a large majority of the grand- 

 children are of the same type as their 

 grandparents, while a small minority show 

 new combinations of the grandparental 

 characters. To change the statement, if 

 A and B enter the hybrid together they 

 tend to come out together in the grand- 

 children ; if they enter separately they 

 tend to come out separately. Why should 

 this be so? 



The facts will become clearer if we look 

 again at the actual case of the fruit-flies 

 already referred to, which was worked out 

 by Professor Morgan. The grandfather 



combines white eyes and yellow body color ; 

 the grandmother red eyes and gray color. 

 White eyes and yellow color here enter the 

 hybrid together, while white eyes and gray 

 color, or red eyes and yellow color enter 

 separately. The hybrids in the first gen- 

 eration all show red eyes and gray color, 

 like the mother. On pairing the hybrids 

 together, all four combinations appear — 

 red eye and gray color, white eye and yel- 

 low color, whit^eye and gray color and 

 red eye and yellow. The last three of 

 these are seen only among the males; for 

 although also present among the females 

 they do not come into actual view, because 

 in this sex white eye or yellow color is 

 dominated or concealed by red eye or gray 

 color. We may therefore confine our at- 

 tention to the males. Now, on counting 

 the relative numbers of these types among 

 the grandsons a remarkable result con- 

 stantly appears. An enormous majority 

 of them — more than 100 to 1 — show the 

 same combinations as the grandparents, 

 namely, white eye and yellow color, or red 

 eye and gray color; while the two new 

 combinations, white eye and gray color, 

 and red eye and yellow color, are corre- 

 spondingly rare. This, I repeat, is obvi- 

 ously because the characters that enter the 

 hybrid together tend to come out together 

 in the grandchildren ; those that enter 

 separately tend to come out separately. 

 This at once suggests that the difference of 

 result depends upon whether the two char- 

 acters in question are borne by a common 

 carrier in the germ-cells or by different 

 carriers. White eye and yellow color tend 

 to hold together because they enter the 

 hybrid in some common carrier. White 

 eye and gray color, or red eye and yellow 

 color tend to remain separate because they 

 enter the hybrid in different carriers. 

 What are these carriers? Very extended 

 experiments, analogous to that just de- 



