396 Ml. G. A. K. Marshall on 



They may be considered as parallel to the Nymi^lialinan 

 among Rhopalocera. In both we have a great preponder- 

 ance of species with cryptic habits and colouring, while 

 genera mimetic of the most distasteful groups of their 

 respective Orders are also common. In some of these 

 examples the mimicry is almost certainly Mlillerian, as in 

 the case of Ncptis and Limcnitis among the Nymphalin<v 

 and Cymatiira and (7cro2?/csis among Longicorns. Further- 

 more, the Clytinx- and GaUichrominaj have been shown by 

 Mr. Shelford to be mimicked in Borneo on a large scale by 

 other Longicorns, although the former tend strongly to 

 mimic Aculeates in nearly all parts of the world. Simi- 

 larly, there is good reason for thinking that the genera 

 Nqjtis and Limcnitis, which may be mimetic, are also 

 themselves mimicked by other Nym^plialinm, etc. 



There remain the remarkable cases of the Ciciiiddidm, 

 Carabidiv,3ind CurculionidcV, which are probably without any 

 strict analogy in the Lepidoptera. The two former require 

 special mention, the last-named are treated separately on 

 pp. 522-525. 



CiCiNDELlD/E. These Colcoptera supply models for 

 mimicry by a Locustid in Borneo, and Mr, Shelford also 

 considers that one of them is mimicked by a fly. They 

 are also known to be mimicked by Longicorns. Many 

 South African CicinddidcV are convergent towards, or 

 mimetic of, CarabidiV, especially tliose of the latter which 

 are themselves convergent towards Mutillidcv(see pp. 511- 

 515). Some of the species have a peculiar scent. On the 

 other hand, their swift movements and retiring habits are 

 inconsistent with a high form of special protection. 



CARABiDyE. Mimicked in Borneo by a Locustid. 

 Probably more strongly defended by the possession of 

 anal glands than are the Cirinddidw, and in the com- 

 binations between the two families it is seen that the 

 latter have approached the former, rather than vice versa. 

 Certain groups of Carahid.v form pseud- or synaposematic 

 combinations with the Galcrucimv and also with the 

 MutillidcV. The South African smaller and moderate- 

 sized diurnal species have habits very similar to the 

 CicinddidiV, but are not so swift. They commonly pos- 

 sess directive marks indicating the specially- protected 

 anal region. The largest South African species {Anthia) 

 have a very large charge of the defensive secretion and 

 extremely powerful mandibles. They freely expose them- 



