DIRECTION AND FREQUENCY OF MUTATION 



219 



the other bars. While it is probable that such intermediate mem- 

 bers of the series may be definitely demonstrated in the futm-e, it 

 is clear that the recorded changes are between a few definite 

 points and not from one point to any other point whatsoever. 

 No such conclusion could have been reached if individual muta- 

 tions had appeared only once or twice. In these observations 

 the change from bar to full came fifty-two times and that from 

 ultra-bar to bar five times. 



A comparison with the white-eye series is interesting. Figure 

 4 gives all the changes observed in that series, according to 

 Miiller ( '20) . The difference between any two steps is so slight 

 that there may be a question as to whether the components of 



.00000 ^^ 



^full*^ 



^ 



■> 



.00003 \/ 



< 



bar 



-^ ultra- ^^^ 

 bar ^ 



/\ 



.00061 



A. 00058 



.00035 



Fig. 3 The bar-eye series of multiple allelomorphs with the direction and 

 frequency of all mutations. Note that the reverse mutations are much more 

 numerous than the direct and that all three possible kinds of jumps in this direc- 

 tion are realized. 



the series as given may not represent the limits of possible dis- 

 crhnination. According to the students of the series, however, 

 there is never a doubt as to the classification of particular indi- 

 viduals, and this would seem to eliminate the view of existence 

 of unperceived intermediates. The conclusion, however, is 

 vitiated by the fact that most of the mutations have been 

 observed only once. 



2. What kinds of jumps occur between the three stopping-places? 

 An examination of figure 3 shows that five of the six possible 

 kinds of jumps have Occurred. The only one that has not been 

 observed is that from full to ultra-bar. When it is considered 

 that the jump from full to bar has not occurred at all in my 

 material, the only recorded case being the original mutation as 



