78 Bulletin American Museum of Natural History [Vol. XLII 



(page 81), based entirely upon larval characters, has been erected. 

 Just what value the larval characters have for purposes of classification 

 is a question. As a rule, the larvae and adults of this family appear to 

 have evolved hand in hand but, in some cases, it is evident that one 

 specialized while the other tended to remain primitive. Nevertheless, 

 some value must be attached to these larval characters. 



Many controversies have arisen over the question of what deter- 

 mines a primitive larva. Brauer, 1869, established the well-known law: 

 "Je naher Larve und Imago einander und der form der Stammkerfe 

 stehen, oder je weniger die imago, die Larve and vollkpmmener Organi- 

 sation ubertrifft, destoalter ist die Form." The essential thought of 

 this law has stood ever since but qualifications have been necessary. 

 Peyerimhoff, 1900, in discussing this law, stated that a careful examina- 

 tion of aU parts is necessary to prove that a larva, apparently campodei- 

 form, is absolutely primitive. 



Lameere, 1899, believed that holometabolism was brought about 

 by the boring of insects in vegetable tissues. As a result, an eruciform 

 type of larva was evolved. In 1903, he said that Brauer's law could 

 only be accepted for insects without metamorphosis and that supposed 

 campodeiform larva? of holometabolic insects are only campodeiform 

 in appearance, being the result of etiological convergence. They are 

 derived from eruciform larvae adapted to feeding on vegetation. The 

 larvae of the primordial holometabolic insects thus acquire these special 

 characters only through the influence of similar habits. 



It was the belief of Gangldauer, 1904, that the Hydroplulidse con- 

 stitute a terminal family of a primitive branch of Coleoptera whose 

 stock has been lost. The natural place, he said, is between the Staphy- 

 linidse, with which they agree in the presence of larval cerci, and the 

 Diver sicornia, with which they agree in venation. He thus indicated 

 that the Diversicornia, which he had established in 1902, was not 

 monophyletic ; and, at the same time, he indirectly accepted Lameere's 

 grouping of the Hydrophilidse under the suborder Palpicornia. This is 

 apparently the most logical place for the family. 



In this paper the true campodeiform larva has been considered as 

 the most primitive type. By true campodeiform larva is meant that 

 type of campodeiform larva which has not been evolved by etiological 

 convergence, but one which is strictly primitive. One finds but little 

 difference between Limnebius, Hydrcena, etc., and the primitive type. A 

 careful examination of the specific parts in these genera reveals only 

 such characters as are fundamentally primitive. Apparent vestiges of 



