March 2, 1922] 



NA TURE 



277 





ember of that group is determined. It is then 

 lecessary to find its position with respect to another 

 own member of the group (preferably one near by) 

 hich determines whether it is north or south of the 

 t member. Once this has been done, the method 

 inheritance of the new character with all other 

 embers of its group can be worked out on paper from 

 e crossing-over data, plotted as distance. In other 

 rds, the heredity of this new mutant, with all the 

 ther known characters of Drosophila, can be predicted, 

 since, with its normal allelomorph, it will give a 3 : i 

 atio ; with any character in another group it will 

 e a 9:3:3:1 ratio ; and with other members 

 its own group it will give a definite result which 

 can be calculated from the " distance " of the 

 plotting. 

 The second point of significance concerning the 

 tting of the genes in terms of distances is as follows : 

 le discovered relation of genes, as expressed in dis- 

 ces, is one that holds for points in a line. This 

 cans that if the genes in question are represented in 

 ce, their relation to each other is that of points in a 

 e. If the line is a chromosome, then the chromo- 

 mes are to be thought of as made up of a single line 

 genes. The reasons for referring the genes to the 

 omosonies have already been given. The possi- 

 bility of explaining crossing-over on 

 0.0 T Yellow a chromosome basis will be discussed 

 later. 

 There is one situation where, on 

 1^ T" White superficial examination of the data, 

 an apparent disturbance of the 

 linear order may appear, namely, 

 when crossing-over takes place at 

 two levels in the same linked series 

 at the same time (double crossing- 

 over)." But by marking intermedi- 

 5.5 + Ec h i n u s ate points between the extreme ones 

 all double cros^-overs can be de- 

 tected and the distances corrected 

 , for them. When this is done, it 



T-^ I Ruby at once becomes apparent that 



the linear order is the correct 

 I iG. 14. arrangement of the genes. In 



fact, far from throwing doubt 

 the linear order, these cases, where double crossing- 

 er occurs,* furnish a strong corroboration of the 

 rrectness of the hypothesis. 



The use of the word " distance " as an expression 

 r the percentage difference in crossing-over values 

 es, unfortunately, lend itself to misunderstanding, 

 'ess one knows just what meaning is attached to 

 s. word when used as defined above. An example 

 1 make this clear. If crossing-over is more likely 

 occur in one region of the linear order than in other 

 ions, the plotted " distances " will be relatively 

 short in comparison with the distances of the 

 ainder of the series. Distance, therefore, must be 

 iderstood in a relative, not in an absolute, sense. 

 e have been aware of the necessity of this restriction 

 from the beginning of our studies of the linear order of 

 the genes, and have warned others of the danger in 

 numerous publications, but apparently without Com- 

 plete success. It has also been shown that the per- 

 centage of crossing-over changes under external 

 (Plough) and internal (Bridges) conditions. As the 



NO. 2731, VOL. 109]' 



female gets older, crossing-over becomes less in some 

 cases, hence the " distances " appear to become less. 

 It has also been shown by Sturtevant that genetic 

 factors may exist that affect the crossing-over in certain 

 regions of the linear series, in one case shortening that 

 region to zero, since all crossing-over is suppressed. 

 But the significance of this result, from our present 

 point of view, is that when the shortening factor is 

 removed (by a definite genetic procedure) the original 

 distance of the genes in this region reappears, and the 

 genes are shown not to have changed their original order. 

 This reassures us that the linear order stands on a firm 

 basis. A recent attack on the theory of the linear 



310 

 361 



430 



/ 



444 



56.5, 

 57.0'' 



65.0 

 68.0 



YELLOW 00 



WHITE 



ECHINUS 



RUBY gQ 



CROSSVNL'SS 140 



VERMIUON 

 MINIATURE 



SABLE 



""garnet 



.FORKED 

 BAR 



46.5 

 52.4 



TRUNCATE 

 STREAK 



ROUCHOID 0.0+ BENT 



0.9 EYELESS 



T 



253 



\ 



25.8 



3S.5 

 42.0 

 45.5 



6 35 

 VESTIGIAL 655 



675 

 LOBE 



720 

 CURVED 



98.5 -- PLEXUS 



103.0 -: 5R0WN 

 1050 * SfECK 



Fig.' 15. 



SEPIA 

 HAIRY 



DICHAETE 

 SCARLET 

 PINK 



SPINELESS 

 CLASS 



DELTA 



HAIRLESS 



EBONY 



WHITE-OCELLI 



95.7 = = MINUTE 



order is based on evidence that shows that " through 

 selection " the distances between certain genes changed. 

 The result really has no bearing on the point, because 

 the order of the genes was not shown to have been 

 affected. Moreover, Sturtevant's case, more thoroughly 

 worked out, shows that where even greater changes 

 of distance had taken place the order of the genes had 

 not changed. 



The plotting of the linear order of the genes in the 

 four chromosomes of Drosophila melanogaster is shown 

 in Fig. 15. The four great groups of linked genes are 

 represented by straight lines with the approximate 

 positions of the genes indicated by short cross-lines. 

 The numbers opposite these cross-lines give the dis- 

 tances from a base chosen as far " north " as possible. 

 The location of some of the genes rests on an immense 



