Orr et al.: Systematics of North Pacific sand lances of the genu s Ammodytes and description of a new species 
137 
A. dubius 
A. americanus 
A. personatus 
A. hexapterus 
A. marinus 
A. tobianus 
A. heian n. sp. 
A. japonicus 
Uranoscopus 
A. dubius 
A. americanus 
A. personatus 
A. hexapterus 
1. marinus 
A. tobianus 
A. heian n. sp. 
A. japonicus 
Uranoscopus 
fP A - 
. 59| 99 1— A hexapterus 
A. marinus 
A. americanus 
personatus 
A. tobianus 
■A. heian n. sp. 
- A. japonicus 
Uranoscopus 
Figure 4 
Trees of consensus haplotypes of species of Ammodytes, 
with Uranoscopus as the out-group taxon: (A) most 
parsimonious tree; (B) maximum likelihood tree; and 
(C) neighbor-joining tree. Numbers on branches indi- 
cate bootstrap support. 
test and, with the exception of gill raker length, all 
differed significantly between 2 or more species (Table 
5). Most significantly, the new species had a smaller 
orbit, wider interorbital space, and longer prepectoral 
process than all other species, and A. personatus had a 
shorter preanal length than all other species (Table 5). 
All species exhibited ontogenetic allometry in orbit size 
and pectoral-fin length; the orbit was proportionately 
longer and deeper and the pectoral fin was longer in 
smaller individuals. In A. japonicus alone, gill rakers 
were proportionally longer in smaller individuals. 
Discriminant function analysis All 3 discriminant func- 
tions obtained in this analysis were significant (Wilks’s 
lambda; P<0.0001). All specimens of the 2 species of 
Ammodytes of the eastern and northern parts of the 
North Pacific Ocean (A. personatus and A. hexapterus ) 
were distinguished from specimens of the 2 species of 
Ammodytes found only in waters off Japan (A. japoni- 
cus and the new species) on the first discriminant axis 
(Fig. 5). Scores were calculated from the following first 
discriminant function equation: 
Di = -0.991 (od) - 0.413 ( iow ) + 0.465 (pb) + 0.379 
(pi) + 0.914 ( cpvl ) + 0.048 (pdl) + 0.423 (ppl) + 0.230 
(gr) - 0.192 (Up) + 0.146 (dp) - 0.270 (ap) + 0.875 
(pcv) + 0.445 (cv) - 62.266, 
where Z)l=the first discriminant score of an individual; 
od=orbit depth divided by HL; iou;=interorbital width 
divided by HL; pfe=pectoral-fin base depth divided by 
HL; p/=pectoral fin length divided by HL; cpu/=caudal 
peduncle ventral length divided by HL; pcP=predorsal 
length divided by HL; pp/=prepectoral length divided 
by HL; gr=the number of gill rakers on the entire gill 
arch; llp= the number of lateral-line pores; dp= the num- 
ber of dermal plicae; ap=the number of anal-fin pteryg- 
iophores anterior to the first haemal spine; pcv= the 
number of precaudal vertebrae; and cu=the number of 
caudal vertebrae. 
A specimen with a score above 0 was identified as 
an individual of an eastern or northern Pacific species 
(either A. personatus or A. hexapterus)', scores below 0 
indicated the specimen was a species of Japan (either 
A. japonicus or the new species). 
On the second discriminant axis, all but 3 specimens 
of A. japonicus and the new species were distinguished 
from each other (95.0%; Fig. 5). Three specimens iden- 
tified genetically as A. japonicus were misclassified 
as the new species (Fig. 5). Among individuals iden- 
tified as A. personatus or A. hexapterus, 98.6% were 
distinguished from each other, the exceptions being a 
specimen genetically identified as A. personatus and 
misclassified as A. hexapterus and 2 specimens geneti- 
cally identified as A. hexapterus but misclassified as 
A. personatus. Scores on the second discriminant axis 
were calculated with the second discriminant function 
equation: 
D 2 = -0.334 (od) - 0.578 (iow) - 2.674 (pb) + 0.818 
(pi) + 0.958 (cpvl) + 0.446 (pdl) - 0.765 (ppl) + 0.031 
(gr) + 0.027 (lip) - 0.007 (dp) - 0.675 (ap) - 0.265 
(pcv) + 0.170 (cv) - 2.458, 
where D2=the first discriminant score of an individual; 
od=orbit depth divided by HL; iou>=interorbital width 
divided by HL; p6=pectoral-fin base depth divided by 
HL; pZ=pectoral fin length divided by HL; cpvl= caudal 
peduncle ventral length divided by HL; pdl=predorsal 
length divided by HL; ppZ=prepectoral length divided 
by HL; gr=the number of gill rakers on the entire gill 
arch; llp= the number of lateral-line pores; dp=the num- 
ber of dermal plicae; ap=the number of anal-fin pteryg- 
iophores anterior to the first haemal spine; pcv= the 
number of precaudal vertebrae; and cu=the number of 
caudal vertebrae. 
Most specimens with a score above 0 were identified 
as either A. hexapterus or A. japonicus', scores below 0 
