354 
PACIFIC SCIENCE, Vol. XII, October, 1958 
55). Palpi of normal form for family, odontus 
unidentate, paradont absent. Tibia with no 
clearly defined ctenidium, but with a group 
of unusually smooth setae anteroventrally, 
near the insertion of the tarsus. Tarsus with a 
single solenidion posteriorly and a variable 
number of eupathidia. 
larva: Scutum with four pairs of setae, 
postscutum with two or four pairs; post- 
scutum nearly as wide as scutum. Coxa I 
with a slender supracoxal seta and two ven- 
tral setae, the medial one of which is greatly 
enlarged and bears a number of long, finger- 
like teeth on the posterior margin. Urstigma 
very large, concealed under the posterior mar- 
gin of coxa I in ventral view, visible only by 
transparency. A single pair of intercoxal setae. 
Lassen za-orga.n present. Palpi with only four 
segments visible in ventral view, but tro- 
chanter present as a very small plate dorsally 
on base of palp (Fig. 40). Palpi oriented so 
that the morphological dorsal line forms the 
lateral margin of the palp as seen in dorsal or 
ventral view. Trochanter and patella devoid of 
setae, femur with a single seta. Tibia with a 
deeply cleft odontus plus three simple setae; 
no specialized paradont present. Tarsus re- 
duced to a small hemispherical cap bearing a 
number of setae of variable size and form, 
but one of these is strongly scythe-shaped. 
Rostrum with only the protorostral and tri- 
torostral setae. Basifemur and telofemur of all 
legs fused. Solenidial formula of patella 
(2-1-1), of tibia (2-2-0), and of tarsus (1-1-0). 
Vestigial setae present on patella I and II and 
tibia 1, famulus on tarsi I and II; companion 
setae absent. Tarsi I and II with three claws 
each, III with anterior and median claws of 
same form as on I and II, but with posterior 
claw reduced to a very short rudiment (at 
least in the type species and the species 
described in this paper). 
CHARACTERS SHOWING INTERSPECIFIC 
VARIATIONS in the genus Paratrombium 
Although the study of interspecific varia- 
tion within a genus is of practical value in the 
identification of species, this should never be 
regarded as the sole purpose of such studies. 
Any type of variation observed, whether of 
practical utility in a key or not, should be 
investigated as it may indicate relationships 
within the genus, or evolutionary trends. 
When combined with similar studies in re- 
lated genera, they provide information on 
whether or not the same types of adaptive 
changes have occurred in different genera. In 
the large Parasitengona, intensive studies of 
variation are even more important than in the 
other genera, because the differences separat- 
ing species are more often of a purely relative 
nature; that is, they involve variations of con- 
tinuous rather than discontinuous types. As 
a consequence of the studies of the two forms 
described here, a number of characters were 
discovered which show variations of possible 
specific value. These are summarized below, 
the characters for adult and larva being listed 
separately. Most of the character variants 
listed in the tables are self explanatory, but 
some require further explanation. Most of 
them show variation of a continuous type. 
One exception is found in the dorsal chaeto- 
taxy of the larva, in which the postscutum has 
two setae in one of the species and four in the 
other. These variants are discontinuous only 
in the sense that no intergradations between 
them have been found up to the present time. 
While it is not inconceivable that these exist, 
it is interesting to note that the closest ap- 
proach is found in occasional specimens of 
P. bidactylus in which there are two setae on 
one-half of the postscutum. Even in these 
cases, the half of the postscutum in which the 
duplication has occurred does not resemble 
its counterpart in P. quadriseta , since the 
placement of the two setae is markedly 
different. 
The variations in form of the hysterosomal 
setae (character 14, adult) are more or less 
discontinuous, although it is quite certain 
that many gradations will be found when the 
genus is more adequately known. If Feider’s 
assumption (1952) that nearly all species of 
