28 THOMPSON YATES AND JOHNSTON LABORATORIES REPORT 
been lost in the Sarcopsyllidae in consequence of the fixing themselves permanently 
on the host by means of the mandibles and upperlip. As already described by 
Enderlein, the labial palpi of the Sarcopsyllidae which are attached to the host, are 
lying on each side of the head and point upwards. We have observed this position 
in Echidnophaga and Hectopsylla. Frauenfeld mistook the labial palpi for pale 
bristles. In all of the Sarcopsyllidae and in the which are unattached to their 
host, the labial palpi are directed downwards and form a sheath for the mandibles and 
upperlip as in other Siphonaptera. 
The great loss of segmentation in the rostrum is doubtless secondary to the 
reduction of the chitinization, as a long and strongly chitinized rostrum with only 
one or two joints in it would offer too much resistance when the piercing organs are 
pushed into the skin. The labial palpi of the Pulicidae consist normally of five seg- 
ments, the number being reduced in true Pulex, while in some American Pulex (for 
instance in Pulex cleophontis) there are six segments. Some stationary Pulicidae like 
Eermipsylla and the Bear-fleas have more segments. This is interesting, for in these 
insects greater segmentation lessens the regidity of the long rostrum, as reduced 
chitinization does in Sarcopsyllidae. The greater number of segments occurs in Vermi- 
psylla. The labial palpi of Malacopsylla consist of five short segments. 
We have already referred cursorily to the great reduction of the thoracic seg- 
ments which is observed in all the Sarcopsyllidae. All the portions of the thorax are 
reduced to a great extent, with the exception of the prosternite and metathoracic 
epimerum. The value for the individual of this contraction of the thorax is obvious. 
The abdomen of a Sarcopsyllid specimen which is fixed on or in the skin of the host 
does not project so much as it would were the thorax of normal length. The parasite 
is therefore less likely to be rubbed off by the host. It might further be argued 
that, as the swollen abdomen of the pregnant $ so presses against the thoracic tergites 
that they telescope into one another, a long thorax would be much in the way. A 
third reason, of a less speculative nature, for the reduction of the thorax, is the loss 
or reduction of the power of jumping, and hence a reduction of the muscles in the 
thorax. In the genus Dermatophilus the thorax is so short that the occiput almost 
touches the abdomen, even in non-pregnant 
The dorsal and ventral parts of the thorax are best considered separately. The 
three thoracic tergites, which are together always shorter than the first abdominal 
tergite, bear only a remnant of the rows of bristles normally present in Siphonaptera. 
In the Sarcopsyllidae each tergite bears a small number of thin bristles placed in a 
verticle row, which row corresponds to the posterior row of bristles present in other 
Siphonaptera. On the mesonotum these bristles are much reduced in number, but 
there are often one or two minute hairs in front of the row, these hairs representing 
another row. This second row is frequently met with in Siphonaptera, but is seldom 
present on the prothorax. The presence of two rows of bristles on the tergites may 
