OF MANDIEULATA. 433 



Hence may be derived the following tribes : 



Termitina Metamorphosis semicompleta 



Corydalina Metamorphosis incompleta - 



Myrmeleonina Metamorphosis obtecta 



Libellulina Metamorphosis subsemicompleta 



Panorpina Metamorphosis 



Although these groups are laid down in the Regne Ani- 

 mal, I have no great confidence in their accuracy on ac- 

 count of our knowing so little of the exotic Neuroptera. 

 The metamorphosis even of the common Panorpa is still 

 unknown; and therefore it is entirely from theory that I 

 conclude it, in the above arrangement, to be intermediate 

 between that of a Termes or Psocus and that of a Libel- 

 lula. If we consider indeed the extraordinary general 

 resemblance which Nemoptera bears to the Ephemeris, and 

 which the Panorpa hyemalis of Linnasus bears to Raphi- 

 dia and Termes, this idea seems not improbable. But 

 even if it should be erroneous, this can never affect the 

 truth of the principal fact we have to establish after the 

 suggestion of M. Latreille, namely, that the essential cha- 

 racter of the ISieuroptera is a varied metamorphosis. Their 

 larvae undergo either incomplete, obtect, subsemicomplete 

 or semicomplete metamorphosis, in opposition to the Or- 

 thoptera,yv\v.c\\ are subject only to semicomplete metamor- 

 phosis, or to the Trichoptera, which have it only obtect. 

 But although the Neuroptera, like their corresponding or- 

 der among the Haustellata, vary so much in their meta- 

 morphosis, the Libellulina, which are tlie types of the 

 order, are subject to one which is peculiar. We proceed, 

 however, with our affinities. 



Linnaeus gives the following singular description of an 



2 F 



