172 MR. H. BOLTON ON INSECT-REMAINS [Feb. 19II, 



and have frankly avowed that, in the present state of our know- 

 ledge, the classification of the Palaeozoic cockroaches is largely 

 artificial and a matter of convenience. That such is the case I 

 believe few will venture to dispute. 



After a considerable amount of comparison and research, I 

 have abandoned the possibility of blattoid affinity, and find the 

 nearest ana]ogues to this wing in the genus Lamproptilia of 

 Charles Brongniart. That author described and figured two species, 

 L. grand' euryi and L. stirrupi, 1 which show a close agreement 

 with the form here described. This is more closely related to 

 L. stirrupi than to L. grand'euryi, it is more quadrangular in 

 outline than either, and the costal area is much broader. The anal 

 portion of the wing is of greater tenuity, and occupies fully half 

 •of the inner margin, while the latter is almost straight in its distal 

 half. To distinguish it as a species, I assign to it the name of 

 L a mp r op til i a tenuit eg m i n at a. 



Horizon. — Level in No. 2 Khondda Seam, 1| miles north-east 

 of Iiesolven Station (Glamorganshire). 



IV. General Observations. 



One feature of some interest is the marked association of these 

 blattoid wings with vegetable remains, especially the leaves of 

 Cordaites and the pinnules of JS'europteris. In several cases the 

 wings are interbedded with Cordaites leaves. The wing of Hemi- 

 mylacris obtusa occurs in a typical underclay, and that of Lampro- 

 ptilia in a yellowish-brown shale crowded with plant-remains. 

 The proximity of the deposits to a land-surface seems evident. I 

 have elsewhere alluded to the presence of sunken pits in the 

 Cordaites leaves, and the possible bearing which these may have 

 upon the use of Spirorbis pusillus as food by the blattoids. Scudder, 

 Sellards, and several others have noticed the association of blattoids 

 with fossil plants, Dr. Sellards suggesting that they were fond of 

 moist low places with abundant vegetation, such as would be found 

 along the banks of rivers and marshes. 2 From such situations the 

 transference of dead insects and loose vegetation into the water 

 would be inevitable. 



Any attempt to discuss the relationship of these South Wales 

 blattoids with those of North America or Europe could only be 

 based upon somewhat vague conjectures. The presence of Archi- 

 mylacrid and Orthomylacrid forms, no less than the presence of 

 Lamproptilia, is indicative of a considerable advance in insect 

 development in this country beyond the more primitive palaao- 

 dictyopteran types ; and from their abundance we may fairly 



1 'Etudes sur le Terrain Houiller de Commentry ' vol. iii (1893) pp. 467-70 

 & pi. xxxv (19), figs. 7-9. 



2 E. H. Sellards, ' Structure of Palaeozoic Cockroaches' Amer. Journ. Sci. 

 ser. 4, vol. xviii (1904) p. 122. 



