425] COLORS OF TIGER BEETLES— SHELF ORD 31 



ication of seven transverse dark bands with six transverse light bauds 

 between them (page 161). The figure of the hypothetical pattern is in 

 general terras almost identical with that shown for Ciciudela and inde- 

 pendently conceived on plate I, figure 4. In this second and third 

 light band are represented by a single wide one and the fifth and sixth 

 are separate as two narrower bands. If the general plan of longitudinal 

 and cross bands in insect patterns is to be accepted we must also con- 

 clude from the evidence presented that the relations to trachea, may be - 

 reversed, i. e., the pigmented areas may lie immediately above the 

 tracheae or between them. In the Lepidoptera pigment appears last 

 in the veins (JIayer, 1896). 



The areas between the trachea may be subdivided into two longi- -W 

 tudinal bands. The pigmented and'unpigmented bauds may also be 

 reversed iu position as would appear to be the case when we compare 

 the usual cieiudelid patterns with those studied by Tower. There is 

 no reason why this should not be the ease as when markings are lost; 

 the pigmentation which results is often heavier than elsewhere (Figs. 

 135 and 136). 



However when one compares the cicindelid ontogeny with the exist- 

 ing patterns of other orders one finds that they show a series of light 

 spots sncli as might easily correspond to the so-eaUed cells or areas 

 divided by longitudinal and cross veins in a primitive insect such as a 

 may-fly. The may-flies, stone flies and many diptera show such an 

 arrangement in some parts of the wing. At least it may be safely con- 

 eluded that a pattern of faint spots is the primitive type in Cieindelidae 

 if one accepts any of the current criteria for primitive forms. 



I start with this type of pattern as "primitive" with a conscious- 

 ness of the fact that it would be possible to proceed in entirely different 

 directions and from entirely different starting points and make out 

 cases of modification in definite direction fully 'as plausible as the ones 

 here presented, provided only the preceding strong evidence is not ac- 

 cepted. 



On this account it may be well to give the reasons for presenting 

 this matter of modification at all. First, it is presented to further 

 establish the contentions already made as to the character of the pattern 

 plan presented ; secondly, to show that all even the most specialized 

 types of patterns coxdd have been derived from the generalized types 

 described above: thirdly, to show that there are certain laws of modifi- 

 cation which nuist have been very general in tiie group and which have 

 operated again and again in the production of the characteristic types 

 of patterns. 



Figures 149&, 156a, and 165, plate XII, show some of the pat- 

 terns in which five nearly complete cross bands occur; 179 shows a very 



