60 ILLINOIS BIOLOGICAL MONOGRAPHS [454 



ual variation the characteristic joining to make the "middle band" is 

 indicated. This occurrence of cross bands as noted and the variations 

 of interrupta together with the light stripe in the region of joining 

 of the cross band 4 with cross band 5.6 which occurs in the ontogeny 

 of the patterns of Scutellaria constitute the evidence for the line of 

 development suggested. 



The second tendency to be noted is the shifting of the spots near 

 the sutural or anal border of the elytron out of line with the cross 

 band with which are properly associated. This is shown in figures 

 156 and 156 a, plate XII, interrupta and in figures 153 and 154 in 

 flexuosa. 



The third tendency to be noted is the loss of the three small baso- 

 sutural spots (Bl, C2.3, D4, Fig. 49, PI. V). This usually takes place 

 in a definite order if individual variation may be trusted as an indi- 

 cator. At least these may have disappeared in some definite order 

 leaving the typical pattern of tranquebarica as shown in the controls 

 of the experiments (Figs 456a", b' and 457 a', V). This type is shown 

 in figures 31, 32, and 33, plate III, and the elements from which 

 it is made are shown with others in figure 49, plate V. As further 

 evidence of the longer persistence of C3.4 see figure 125, plate X, and 

 figure 145, plate XI, which are late stages showing the persistence of 

 this spot after the more anterior one has disappeared. 



The fourth tendency which may be noted is the tendency for the 

 typical C. tranqueoarica pattern to shift as indicated in the patterns 

 which result from experimental stimulation during ontogeny. This is 

 shown in figures 456 a, b, 457 a, b, 458, 459, and 460 a, b, plate XXVIII. 

 These modifications have already been noted on page 39 but may be 

 recalled briefly as follows: the forward and backward extensions of 

 the inner end of the humeral lunule (spot B2 drops out) disappear; 

 the slight forward extension of the inner end of the middle band in 

 the longitudinal stripe C(C5) drops out or loses identity. The with- 

 drawal of the middle band from the elytral margin and reduction to 

 conform with that of C. purpurea (purpurea) (Fig. 537, PL XXVIII) 

 is the striking and probably the most important change best illustrated 

 in 460 a, b,. Similar modifications in all high temperature experiments 

 with C. hirticollis (some with C. limbalis) serve to clinch the argument 

 for response in definite directions. 



A fifth tendency is illustrated by C. purpurea as shown in the 

 figures to the left in figure 537, plate XXVIII. The differences be- 

 tween the purpurea series and the tranquebarica series lies in the short 

 humeral lunule of the former, which indicates a different tendency 

 which perhaps constituted the original distinction between the patterns 

 of the two series. 



