SEGMENTATION AND PHYLOGENY OF ARTHROPODA. 479 
The basal sclerite of the palp-bearing maxilla in Polyxenus 
agrees with the external stipes of the Julid gnathochilarium. 
These are the first pair of maxille, and in Polyxenus they 
lie for the most part in front of, not exteriorly to, the second 
maxilla (labium). Thus we see the specialised diplopodan 
structure in process of formation from two pairs of jaws. 
In the relative positions of maxille and labium, Polyxenus 
is more primitive than Scolopendrella, in their partial fusion 
more specialised. 
When the hind head skeleton of Polyxenus is viewed from 
within, we find that the flat plate seen and figured by 
Latzel, consists not of the tongue only but of that organ 
(figs. 1, 2, 4 ld) together with a pair of maxillule (figs. 
1-5, mal). The outer edge of these jaws can be clearly seen 
from behind (fig. 1), and when the maxillula is viewed from 
within (fig. 2) or isolated (fig. 3) its likeness to the corres- 
ponding appendage in Scolopendrella or a Springtail is 
unmistakable. It is a delicate, transparent, chitinous plate 
with outer and inner lobes, each with two or three promin- 
ences and the inner with several bristles at the tip. On 
its posterior face, close to the tip, the maxillula bears a long 
flagellate process (figs. 1, 3, 4 fl.) which assumes various posi- 
tions with reference to the other structures (cf. figs. 1, 3) 
in different individuals which I have studied. ‘The inner lobe 
of the maxillula is in close contact with the tongue, and the 
lateral region of the tongue unites with the hinder face of the 
maxillula near its outer edge (figs. 1, 2 1). 
When we remember that the maxillular segment has only 
once been detected in an embryo insect, it is not surprising 
that it has hitherto been overlooked by the few zoologists 
who have studied the development of millipedes. The general 
agreement of embryologists that the diplopodan gnathochi- 
larium arises from one pair only of maxillary appendages 
is, however, a more serious objection to the interpretation of 
the structure which I, following Hansen (6), advocate here. 
But Heymons (10), in his account of the germ-band of Julus 
and Glomeris, describes a post-maxillary segment which he 
