26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I40 



flattened (fig. 5, A, indicated by broken arrows) the resulting brain 

 would resemble that found in the geophil head capsule. The free 

 frontal ganglion and the stomatogastric ganglion are now incorpo- 

 rated in the neural mass instead of having their own neurilemma. 

 The antennal lobes are joined at the midline while still separated by 

 a cellular cortex, and the protocerebrum is compressed. Figure 5, B 

 represents a schematic sagittal section through the supraesophageal 

 ganglion of hipuncticeps. Terms can now be assigned, tentatively, to 

 the various glomerular elements in the brain of hipuncticeps. 



The deutocerebrum is considered first. About 60 micra posterior to 

 the base of the antennal nerves, a marked indentation of the neuropile, 

 approximately 40 micra in length, exists. This suggests the presence 

 of a medial and a lateral lobe (pi. 2, fig. 3, Dl, Dm). Between these 

 lobes courses a fibrous structure which is probably homologous to the 

 "corpus lamellosum" (elm) found in the other three orders of chilo- 

 pods. As the name suggests, this glomerulus consists of densely ag- 

 gregated parallel fibers. This tract consists of fibers which run be- 

 tween the medial and lateral lobes in a dorsolateral-ventromedial 

 direction. The tract bends abruptly caudad about 100 micra from the 

 anterior margin of the brain. It continues into the circumesophageal 

 connectives. It receives fibers from the cells of the "pars intercere- 

 bralis" (pi. 3, fig. i). Whether the tract of the opposite side sends 

 fibrous connections by way of the region of the stomatogastric bridge 

 is not certain. The fibers lose optical individuality in this region and 

 cannot be followed with precision. 



A bundle of fibers in the lateral lobe of the deutocerebrum follows 

 a path parallel to the tract just described. It assumes a position lateral 

 to the corpus lamellosum in the circumesophageal connectives. These 

 two pathways undoubtedly establish connections between the antennae 

 and the subesophageal ganglion. Whether they are motor or sensory 

 or both cannot be determined in the present study. The evidence that 

 connections exist is indirect and is suggested by the manner in which 

 this species "cleans" its antennae with its mouth parts. 



In addition to a large glomerulus ventrally situated at the rostral 

 limit of the antennal lobe (pi. 2, fig. i, glom), two small adjacent 

 dorsolateral glomeruli are present in the medial lobe of the deuto- 

 cerebrum. These are most probably the "antennal glomeruli" de- 

 scribed by other authors. In a brain of this size they may be easily 

 overlooked. Saint Remy mentioned the presence of "glomerular con- 

 centrations" in N ecrophloeophagus longicornis but gave little attention 

 to their histology. The antennal glomeruli in hipuncticeps lack the 

 scalloped appearance of the antennal glomeruli in the higher arthro- 



