122 PHYSICAL BASIS OF HEREDITY 



never more tlian 50 per cent., altliough the actual num- 

 bers given for ^'distances'' between two genes may be as 

 much as 107 based on summation of short distances. The 

 latter method of calculation is the accurate way of stating 

 the result, and whenever possible it is adhered to, i.e., 

 the percentage numbers for crossing over are sum totals 

 based on results obtained with genes so near together that 

 double crossing over is practically excluded. 



Another illustration where the ditference between the 

 direct calculation between two factors (scute and forked) 

 and the '^ piece-by-piece " estimate is greater than 50, is as 

 follows : At one end of the series of sex-linked genes is a 

 factor scute (zero) and near the other end forked. The 

 direct data for crossing over between them gave a cross- 

 over value of 48.2. Between them three other loci were 

 present in the same experiment, and crossing over between 

 them could also be detected. As shown in the table below, 

 the sum of these crossover values gave 61.1 units between 

 scute and forked. 



Scute 



48.2 

 Garnet 



14.2 



Forked 



The presence of the intermediate factors makes it pos- 

 sible to pick up most of the double crossing over that 

 occurred between scute and forked. When a correcticm 

 is made for these the diiference between 48.2 and 61.1 

 entirely disappears. Another and still more extreme 

 example will help to make this more evident. Near one 

 end of the second chromosome is the gene for star (eyes), 



