BLATTOIDEA. 95 



egg-capsule. Cerci are present, and Sellards has identified long ovipositors, 

 though this identification is not accepted by Handlirsch. 



The habits of Carboniferous Blattoids can only be inferred from those of living 

 forms. I have previously (p. 11) drawn attention to the occurrence of Blattoid 

 wings among the leaves of Gordaites bearing numerous shells of the small annelid, 

 Spirorbis piisillvs, and suggested that the Blattoids frequented decaying vegetation 

 to feed on the Spirorbis and similar organisms. 



The tegmina of Phylomylacris mantidioides occur among masses of the spat 

 ("Ancylus vintii") of Anthracomya phillipsii in coal shales of the Durham Coalfield. 

 Most of the other Blattoid remains from the Coal Measures of this country have 

 been found in sedimentary deposits in which Ostracods also occur. Apart from 

 these facts, the general faunal associations of fossil insects to which I drew attention 

 on pp. 18 24 are, I think, of great importance, as furnishing definite clues to habits. 

 The same Arthropod-association still persists, and the lagunal and marshy phases 

 of the Coal Measures would furnish a fitting and desirable environment for the 

 whole series. 



Classification. The classification of the Carboniferous Blattoids is difficult, 

 owing to lack of material. The known British forms are few in number, and 

 have been found in several coalfields at horizons which cannot be correlated with 

 each other or placed in a regular sequence. 



Scndder's classification does not sufficiently take into account the relationship 

 which undoubtedly exists between the Blattoids and the Palasodictyoptera, and it 

 is based mainly on a knowledge of living forms. For this reason, probably, the 

 Mylacridae are placed as the first tribe in his group " Palaeoblattariae," in 

 preference to the Archimylacridas, in which a Palseodictyopteroid relationship is 

 more clearly evident. 



Handlirsch's classification gives greater attention to the ancestral type, and is 

 also based on an examination of more numerous specimens than were accessible to 

 Scudder. It is far from satisfactory, and will certainly be much modified when 

 the collection of several thousand Blattoid remains from Commentry is fully 

 worked out by Prof. Meunier, whose results are not yet published. Lameere's 

 general classification ('Bull. Mus. Hist. Naturelle/ 1917, no. 1) is too uncertain 

 to justify adoption. 



Dr. P. Pruvost, who has given several years to a close study of the fossil 

 Blattoids of the French and Belgian Coalfields, has modified Handlirsch's classifica- 

 tion, and in his latest memoir (1920) he classifies the Blattoidea of Northern 

 France as follows : 



BLATTOIDEA. 



Family ARCHIMYLACKID.S;. 



Genera ActinoUatta, Manoblatta, ArcUmylacris, Asemollatta, PhyloUatta, Grypoblattina, 

 Arcluvotiphe, Barroisiblatta, and Mesitoblatta. 



