ORTIIOPTERA AND NEUROPTERA. 619 



terous larva. The adult wing-venation is unusual. It is, I am 

 sure, most incorrect to " lump " this highly specialised group with 

 any others. The Odonata are justified in being given ordinal 

 rank. 



Sharp "lumps," in his suborder Amphibiotica of the extended 

 Neuroptera, the Odonata, Perlidae, and Ephemeridae. The 

 aquatic larval life appears to me as most obviously an indepen- 

 dently-acquired adaptation in each case. The entire difference in 

 means of aquatic respiration between these groups is strong 

 evidence in point. The venation in adult wings is also evidence 

 against uniting these groups. The Odonate wing, with its crossed 

 medial and radial veins, dynamically specialised nodus, anal 

 region, etc.. is in strong contrast to the primitive Perlid wing and 

 the almost " orthopterous " Ephcmerid wing. 



Crampton has shown, in the case of the Perlidae, a strong 

 resemblance to various Aptera in many features. This primitive 

 adult-form, coupled with a sharply-defined two-stage metamor- 

 phosis, may well justify us in adhering to this group's name as the 

 order Plecoptera, following Comstock. 



The Ephemeridae are quite as distinct from the Plecoptera 

 as from the Odonata. To save time. I shall not give the 

 evidence here, but shall simply rank thenl as the order 

 Ephcnierida, as does Lang. 



One constituent of the l^innean Neuroptera is the family 

 Terniitidae. An inspection of our adult winged termites in any 

 African museum will show others, as it has shown me, the falsity 

 of this classification. A detailed inspection of the various parts 

 of the body reveals mouth-parts, genitalia, etc., much after the 

 pattern of the Orthopterous Blattidae. Were it not for the high 

 specialisation of the Terniitidae, we might well group them into 

 one order with the Blattidae. But the caste-forms of the Termi- 

 tidae prevent this. Hence, I uphold the classification of these as 

 Isoptera, following Comstock and Lang. The name is well 

 chosen, for in no other group, save possibly a few Odonata of the 

 family Caloptcrygidae, are the fore and hind wings so exactly 

 alike. 



Closely allied to the Blattidae is another group, the Mantidae. 

 If the Mantis pidae may be grouped in the Neuroptera along with 

 the Myrmeleonidac, we cannot justly split ofif the Mantidae as a 

 separate order, because they also possess grasping forelegs. The 

 ether characters are quite on the Blaltid model ; the two groups 

 are obviously in one order. 



Enderlen, 1903, divides the OrtJioptera into three groups. 

 Oiie, OotJiecavia, includes Blattidae and Mantidae. His other 

 groups, Saltatoria and -Gressoria, depend simply upon whether 

 these Orthoptera walk or jump. This. I submit, is merely an 

 adaptive character, produced by the elongation of the hind-legs, 

 which are the longest pair in all groups of the lower winged 

 insects. Accordingly, I should prefer a somewhat djfferent 

 classification. 



Between Gryllidae, Locustidae, and Acrididae there is much 



