FOSSIL COCKROACHES OF NORTH AMERICA. 183 



in temperate Europe, here in Belgium itself, the zoologist who is really an observer meets at 

 every step cases of dissimulation which are every whit as striking as those which tropical 

 nature offers us." ' 



The arguments I have used elsewhere in discussing this subject - attempt to show that 

 in the very nature of things protective resemblance must prevail in a world where creatures 

 are the food of others, and escape destruction when observed by their predaceous foes less 

 easily or less frequently than their fellows. From this standpoint it would be difficult to 

 refrain from the logical conclusion that protective resemblance was nearly or quite as much 

 a feature of past life as of present. 



Naturally, since colouring forms the next important or the most common part of pro- 

 tection, proof of such protection cannot be derived from the fossils. But pattern of markings 

 is also a conspicuous element of protection in existing types, and in a few fossils among 

 insects we can detect markings of a precisely similar nature to some which in existing 

 insects can be proved protective ; but here habit and association are often necessary factors 

 and these can usually only be inferred in the extinct types, but inferred in some instances 

 with considerable reasonableness. 



The examples which I have in mind are all drawn from Tertiary faunas ; but the reason 

 I refer to the matter here is that it seems to me fairly reasonable to look upon some forms 

 of Carboniferous cockroaches, if not indeed most of them, as probably imitative, and thereby 

 protected. The lirst cockroach wing ever described from the coal was at first regarded as a 

 fern leaf, and in all or nearly all the localities where their remains have been found they are 

 associated with fern leaves in immense abundance. While searching for them in the Per- 

 mian deposits at Cassville, W. Va., I was much struck by their resemblance to each other 

 and was repeatedly obliged to use the glass to determine whether it was the wing of a cock- 

 roach or the pinna of a fern like Neuropteris I had uncovered, and the instances are not rare 

 where they agree completely in size. The general distribution of the nervures is to cursory 

 view the same in each and the contour is often nearly identical. Only the differentiation of 

 the anal area in the cockroach wing at once distinguishes them, but this is I'eally a feeble 

 point and would often be noticed only by an expert. Is it not then plausible to suppose 

 that the intimacy of the resemblance is due, as such an instance of associated organisms 

 would now be regarded as due if the colour agreed, to the action of natural selection in pro- 

 ducing protective resemblance ? The ordinary colour of the fore wings of existing cockroaches 

 is brown or testaceous, yet there are not wanting numerous examples, at least in the tropics, 

 where they are as green as the leaves of ordinary vegetation. 



• Bull. Acad. Roy. Belg-, (3) xxiii., 92. 



2 Atl. Monthly, Feb., 1889; Butt. East. U. S. and Canada 710-720. 



Sec. IV., 1894. 20. 



