Iv. Jordan and N. C. Rothschild 
19 
protruding upwards when the head is embedded in it, as is the case in 
the Sarcopsyllidae (Jordan and Rothschild, 1906, p. 27). The labial palpi 
therefore require to be flexible and this is attained by segmentation or 
by a reduction in chitinization. The latter development obtains in the 
Sarcopsyllidae, all species of which family have a very pale, weak, soft 
rostrum. The necessary flexibility of a prolonged and strongly chitinized 
rostrum as it is found in various fleas ( Vermipsylla , Macropsylla, etc.) 
is obtained by greater segmentation. The species of Loemopsylla, Pulex 
and Pariodontis and the combed fleas Ctenocephalus, Hoplopsyllus, 
Spilopsyllus, and a few forms provisionally placed in Ceratophyllus, as 
well as a number of eyeless fleas, have four segments in the labial palpus. 
All the American non-combed eyed Pulicidae, however, with one excep¬ 
tion, have at least five segments in the labial palpus, the same being also 
the case in the Transcaspian species lamellifer Wagn., which is in other 
respects also closely allied to the American forms. The reduction in the 
number of the segments of the rostrum is not always accompanied by a 
shortening of the orgarj. The rostrum of Hoplopsyllus glacialis, Loemo¬ 
psylla lovgispinus and others, with four labial palpal segments, is longer 
than the rostrum of some fleas in which the labial palpus has five segments. 
The reduction does not take place from the distal end, the apical segment 
often remaining very long in the 4-segmented labial palpus. The 
reduction is not effected by the loss of a segment, but the disappearance 
of an incision, two segments being fused to form one. As the segments 
bear bristles at the apical edge, an intermediate stage of development 
between a 4- and a 5-segmented palpus would be a 4-segmented palpus 
with one of the segments bearing a bristle on each side in or near the 
centre. Such a labial palpus may exist, but we have not yet observed it. 
We think it likely that in the 4-segmented palpus the second segment 
has become fused with the first. 
The bristles at the extreme tip of the rostrum are apparently 
sensory in character like those at the apex of the maxillary palpus. 
The insect is probably using them to test the skin of the host prior 
to puncturing it, at least so it appears when a hungry flea is placed 
on the hand or arm. These “testing” bristles of the rostrum differ 
both in number and size in the various groups of fleas; and as the 
variation within each group is but slight, they afford characteristics 
which are valuable for classification. There are apparently never more 
than six such bristles at the tip of each labial palpus, this being, it would 
seem, the normal or ancestral number, three bristles being placed on each 
side at the apical edge of the end-segment. This number is found in 
2—2 
