12 MISC. PUBLICATION 5 00, U. S. DEPT. OF AGRICULTURE 



Crampton assigns to the fleas the rank of a superorder, it being 

 one of the three superorders into which holometabolous insects are 

 divided. His arrangement of these superorders follows: The 

 Pansiphonaptera, including the Siphonaptera only; the Panmecop- 

 tera, including the Neuroptera, Mecoptera, Diptera, Lepidoptera, and 

 Trichoptera, with their fossil allies ; the Pancoleoptera, including the 

 Coleoptera, Strepsiptera, and Hymenoptera, with their fossil allies 

 except the Protohymenoptera. 



The theory that fleas have been derived from the Trichoptera was 

 assailed by Tilly ard (74, p. 39) who stated that— 



There is not a shred of evidence in favor of Fleas having been derived from 

 an aquatic ancestral type. . . . For Fleas to have been derived from Tri- 

 choptera, it would be necessary to postulate the loss of the ancestral functional 

 mandibles and the evolution of the present type of elongated spear-like organ 

 from the nonfunctional vestige of the Trichoptera! This also involves a belief 

 in the loss and subsequent reappearance of the mandibular muscles, and contra- 

 venes Dollo's Law.. 



* * * Fleas are smooth and shiny, very unlike the soft-haired Trichoptera in 

 their body-covering. What hairs they possess are specialized into stiff bristles 

 and combs. 



Tillyard (74) claimed that the ancestors of the Siphonaptera must 

 have possessed short antennae, elongated mandibles, four-segmented 

 maxillary palpi, a smooth and leathery body, a metathorax not highly 

 reduced, middle and hind coxae with a meron, and a terrestrial larva 

 feeding on animal or vegetable debris. The pupa must have been 

 a pupa libera, either in a cocoon or a primitive earthen cell. He 

 believed that the only ancestors that meet these conditions are — 



(a) a small, reduced type of primitive Mecopteron, probably of the Upper 

 Permian family Permochoristidae, or (b) a related form classifiable definitely 

 within the Paratrichoptera. 



The objections of Crampton (7) and Tillyard (74) to deriving 

 fleas from ancestral Diptera because of the absence of a meron from 

 the third coxa of the latter is not well founded. Many Diptera have 

 the second coxa with a well developed meron, although both coxa and 

 its meron may be partly or largely transferred to the body wall. 

 Such is the case in the Tabanidae. Other Diptera have the third 

 coxa well developed, with the evidence indicating that there has been 

 a reduction and fusion of the eucoxa and the meron. 



A condition somewhat similar to this is found in certain fleas, as, 

 for example, those of the genus Echidnophaga. In fleas of this genus 

 the meron of the mesocoxa is reduced but separated from the eucoxa 

 by a groove, while that of the metacoxa is partly fused with the 

 eucoxa. Further, it is to be noted that the third coxa of some phorids 

 is not very different in its major aspects from the third coxa of cer- 

 tain fleas, and although there is no suture or ridge dividing it into a 

 true eucoxa and a meron, there exists an undifferentiated area which 

 probably represents the meron. 



The writers believe that much printed space has been wasted dis- 

 cussing the value of the labial palpus as a structure indicating the 

 origin of the Siphonaptera. The segmentation of these appendages 

 is such that they could have been evolved from those of almost any 

 group of free-living insects. The most common number of true seg- 

 ments in the labial palpus of a siphonapteron is five. Yet there are 

 several genera with four, three, or two segments in each palpus. 



