462 WALTER W. MARSHALL AND HERMANN J. MULLER 



abnormality in the region of the cross-vein which connects the 

 fourth and fifth longitudinal veins. 



2. If the modification was merely a bend in the cross-vein, 

 it was classed as 'slight.' Likewise if the disturbance took the 

 form of a slight projection, either proximal or distal, it was 

 considered in the same class, or if the extra mark was appar- 

 ently disconnected or removed from the cross-vein. 



3. If short projections or markings appeared, both proximally 

 and distally, the wing was classed as 'moderate.' Also if only 

 one projection occurred and this was more extended than those 

 of the class 'slight,' even so much as to become branched, it was 

 recorded in this grade. 



4. If fairly extended projections in both directions from the 

 cross-vein appeared, the wings were graded as 'marked.' Such 

 projections were frequently branched, especially distally. Wings 

 in which the venation in this region was confused or considerably 

 disturbed were classed in this lot. 



5. If a small watery blister appeared in this region of 

 modification which was considerably less than half as long as 

 the entire wing, it was placed in the class 'small blister.' 



6. When the bhster was approximately half as long as the 

 wing, it was classed as 'medium blister.' 



7. The grade 'large bhster' was established for a similar con- 

 dition that occupied the entire or nearly all of the wing surface. 



The type and the range of each class being fixed, the question 

 next occurred as to what their order or relative degree of devia- 

 tion from the normal might be. After a comparison of the 

 amount of modification in each class, the order in which they 

 have been described was adhered to, and experimental support 

 for it was also found and will be described presently. 



The method of observing and recording the variations was 

 carefully considered so as to avoid bias. The flies of either the 

 stock (of homozygous ancestry) or of the outcrossed type (of 

 heterozygous ancestry) were inconspicuously marked by ampu- 

 tating a small portion of one member of the front pair of legs. 

 The front pair were chosen because these legs usually project 

 backwards beneath the body rendering them less obvious, and 



