An Interpretation of the Relationships Among the Species 
of Parapercis, Family Mugiloididae 
George E. Cantwell 1 
This work is based on data taken from the 
descriptive study by Cantwell (1964) of the 
genus Parapercis, fishes of the Indo-Pacific 
oceans. The descriptions were based on ana- 
tomical studies, and each structure was analyzed 
to determine its variation within a species and 
its value in identification. 
Here an effort is made to employ those char- 
acters showing the least variation within species 
to establish possible affinities between species, 
to define species groups, and to determine rela- 
tionships among them. 
METHODS 
Using the method described by Cain and Har- 
rison (1958), seven characters were utilized to 
determine the affinity of the members of this 
genus. The characters used were the number of 
teeth in the outer row of the lower jaw, dorsal 
spines, dorsal rays, total anal rays, caudal verte- 
brae, the shape of the spinous dorsal, and the 
connection between the spinous and soft dorsal. 
The latter two characters, not being measure- 
ments or meristic data, were assigned numerical 
values: 5 was given if the dorsal spines became 
progressively longer posteriorly; 4 if the middle 
spines were longest; 3 if the membranes were 
attached to the first soft dorsal ray opposite the 
tip of the last spine; and 2 if a connection oc- 
curred at the base of the first soft dorsal ray. 
Briefly, this method obtains a combined value 
for the differences between two forms. An ex- 
ample of this method is given (Tables 1 and 
2), using only 5 of the 26 species. First, the 
means of the measurements of the same charac- 
ters in all the forms were obtained. Next, all 
the mean values for each character were divided 
by the maximum mean value for that character; 
this gives the reduced values and removes the 
1 Insect Pathology Laboratory, Beltsville, Maryland. 
Manuscript received September 23, 1964. 
bias produced by the different units of meas- 
urement ( Table 1 ) . Third, all the forms are 
compared with each other, two at a time, by 
obtaining the differences between the reduced 
vaues of each character (Table 2). Low values 
imply close affinity, higher values more distant 
relationship. 
When the species are compared with each 
other, two at a time (Table 2), the total dif- 
ferences fall into two categories. Species A, B, 
and C are very much alike, and are different 
from D and E, which resemble each other 
closely. 
RESULTS 
Comparisons of all 26 species of the genus, 
using this method, show the presence of six 
groups of species with great affinity toward each 
other. The mean differences of reduced values 
within these groups are: I, 18.8; II, 10.0; III, 
3.2; IV, 21.7; V, 0.0; and VI, 6.4. In every in- 
stance each member of a group has a greater 
affinity for the other members of its own group 
than for any member of any other group. These 
groups are listed below. 
Group I: binivirgata, multi] as data, mima- 
seana, sexfasciata, muronis, aurantiaca. 
GROUP II: cylindrica, haackei, ommatura, 
pulchella, snyderi. 
Group III: emery ana, filament osa, nebulosa, 
schauinslandi, alb o guttata. 
GROUP IV: cephalopunctata, tetracantha, 
xanthozona, hexophthalma, clathrata, poly op h- 
thalma. 
Group V: ramsayi. 
GROUP VI: colias, gilliesi, allporti. 
The mean differences of reduced values be- 
tween groups give the degree of relationship 
among groups, as shown in Table 3- 
189 
