STURTEVANT 



77 



POSSIBLE OBJECTIONS 

 TO THESE RESULTS 



It will be noted that there appears 

 to be some variation in coupling- 

 strength. Thus, I found (CO)R to be 

 36.7; Morgan and Cattell obtained the 

 result 33.9; for OR I got 34.0, and for 

 CR, 28.5. The standard error for the 

 difference between (CO)R (all fig- 

 ures) and CR is 1.84 per cent, which 

 means that a difference of 5.5 per 

 cent is probably significant (Yule '11, 

 p. 264). The observed difference is 

 6.1 per cent, showing that there is 

 some complication present. Similarly, 

 BiM gave 37.6, while OAl gave 54.0- 

 and BOAI gave 36.7 for BM, and 36.5 

 for OM. There is obviously some 

 complication in these cases, but I am 

 inclined to think that the disturbing 

 factor discussed below (viability) will 

 explain this. However, experiments 

 are now under way to test the effect 

 of certain external conditions on cou- 

 pling strength. It will be seen that on 

 the whole when large numbers are 

 obtained in different experiments and 

 are averaged, a fairly consistent 

 scheme results. Final judgment on this 

 matter must, however, be withheld 

 until the subject can be followed up 

 by further experiments. 



Another point which should be con- 

 sidered in this connection is the effect 

 of differences in viability. In the case 

 of P and M, used above as an illustra- 

 tion, the rudimentary winged flies are 

 much less likely to develop than are 

 the longs. Now if the viability of red 

 and vermilion is different, then the 

 longs do not give a fair measure of 

 the linkage, and the rudimentaries, 

 being present in such small numbers, 

 do not even up the matter. It is prob- 

 able that there is no serious error due 

 to this cause except in the case of rudi- 

 mentary crosses, since the two sides 

 will tend to even up, unless one is very 



much less viable than the other, and 

 this is true only in the case of rudi- 

 mentary. It is worth noting that the 

 only serious disagreements between 

 observation and calculation occur in 

 the case of rudimentary crosses (BM, 

 and (CO)M). Certain data of Mor- 

 gan's now in print, and further work 

 already planned, will probably throw 

 considerable light on the question of 

 the position and behavior of this fac- 

 tor M. 



SUMMARY 



It has been found possible to ar- 

 range six sex-linked factors in Dro- 

 sophila in a linear series, using the 

 number of cross-overs per 100 cases 

 as an index of the distance between any 

 two factors. This scheme gives con- 

 sistent results, in the main. 



A source of error in predicting the 

 strength of association between un- 

 tried factors is found in double cross- 

 ing over. The occurrence of this phe- 

 nomenon is demonstrated, and it is 

 shown not to occur as often as would 

 be expected from a purely mathemat- 

 ical point of view, but the conditions 

 governing its frequency are as yet not 

 worked out. 



These results are explained on the 

 basis of A4organ's application of Jans- 

 sens' chiasmatype hypothesis to asso- 

 ciative inheritance. They form a new 

 argument in favor of the chromo- 

 some view of inheritance, since they 

 strongly indicate that the factors in- 

 vestigated are arranged in a linear 

 series, at least mathematically. 



LITERATURE CITED 



Boveri, T., 1902 "Ueber mehrpolige Mitosen 

 als iMittel zur Analyse des Zellkems." 

 Verb, Phys.-Med. Ges Wurzbiirg., N.F., 

 Bd. 35, p. 67. 



Dexter, J. S., 1912 On coupling of certain 

 sex-linked characters in Drosophila. Biol. 

 Bull, vol. 23, p. 183. 



