ZOOLOGY AND BOTANY, MICROSCOPY, ETC. 
27 
paired ventral saccules, very like those of Scolopendrella, are found at 
the hinder margin of the second to the seventh ventral plate of the 
abdomen ; they are traversed by muscles and a cord of connective tissue. 
No distinct pores can be made out in the cuticle, and the matrix-layer 
is provided with a few gigantic nuclei. In Japyx solifagus, saccules 
beset with glandular hairs are to be seen on the hinder margin of the 
first ventral plate of the abdomen ; in J. gigas these break up into several 
parts. In Machilis there are seven pairs of abdominal saccules with 
well-developed muscles and an apparently non-porous cuticle. The 
ventral tube of Cullemhola on the first abdominal segment has well- 
developed retractor muscles, a cord of connective tissue, and glandular 
cells with distinct pores. Living examples of 31achilis and Fodura show 
that the abdominal sacs are only protruded when the animal is com- 
pletely at rest and in warm damp air. Podurse creeping on a glass 
surface keep their ventral tube generally inactive, and the same is 
always the case with the abdominal saccules of Machilis. 
I'he relation between the development of the tracheal system and 
the ventral saccules shows that the latter have a respiratory function, 
and are to be regarded as blood-gills ; the air-tubes are absent in most 
of the Poduridm, are short and open by a single pair of stigmata in 
Smyntkurus and Scolopendrella, and by three in Campodea ; when the 
common longitudinal trunks are developed the ventral saccules are 
reduced. 
No urinary products can be detected in the abdominal saccules ; 
the development of the saccules is affected not only by the tracheae but 
also by the amount of metabolism which goes on. 
It is probable that the coxal sacs found in both sexes of Diplopods 
have a subsidiary function as organs of attachment during copulation. 
The temporary vesicular appendages of Gryllotalpa, Melolontha, and 
others, the structure of which completely resembles that of the abdominal 
saccules, are to be regarded as secondary blood-gills. 
The stump-like ventral appendages of Scolopendrella (coxal styles) 
are found on the second to the tw'elfth segments of the trunk ; they are 
movable and are traversed by a nerve. The spinning styles of Scolo- 
pendrella are quite immovable, and corresjDond to the cerci of Insects. 
The pair of appendages on the first abdominal segment of Campodea is 
to be regarded as a rudimentary pair of legs ; this genus has no gona- 
pophyses in either sex. The jointed anal cerci of Campodea are quite 
like antennae in structure, but they have no muscles. The abdominal 
styles of Machilis have flexor muscles in the anterior segments ; the 
mid-tail of this genus corresponds to a supra-anal prolongation of the 
anal piece. 
The abdominal styles serve especially as tactile organs and for the 
support of the body in locomotion or in springing, while the anal cerci 
have a function similar to, though less well developed than that of the 
antennae. 
The author concludes that the Myriopoda and Insects have a common 
origin ; the Symphyla stand nearest to the Diploj^oda, but the Pauro- 
poda are to be regarded as degraded from the latter. The common 
ancestors of the Chilopoda and Insecta stand equally near to the former 
and to the Symphyla, but they possessed a posterior genital orifice. 
