NEW AMEIRIDAE FROM W. AUSTRALIAN GROUNDWATER 
and the P5. Males lack a defined P6 closing off the single genital 
aperture and have an extraordinarily large spermatophore. Females 
similarly display a highly reduced genital field. The size of the 
spermatophore, occupying nearly half of the body, is remarkable. 
Except for representatives of the genus Apodopsyllus Kunz 
(Paramesochridae) where similarly sized spermatophores have been 
reported, male harpacticoid copepods produce small spermatophores, 
typically not exceeding the length of two body somites. Given the 
highly reduced genital apertures of /. humphreysi it is difficult to 
imagine how the spermatophore can be successfully extruded and 
transferred to the female. 
Taxonomy of freshwater Ameiridae 
The primary taxonomic subdivisions in freshwater Ameiridae have 
traditionally been based on swimming leg segmentation (Lang, 
1965; Petkovski, 1976) and have ignored other, more phylo- 
genetically informative characters such as setation patterns and 
mouthpart morphology (Galassi et al., 1999). This simplistic 
approach has led to: (1) the generally unsatisfactory practice of 
describing new species virtually exclusively on leg characters with- 
out consideration of cephalic appendages, genital field morphology 
or even female abdominal segmentation; (2) the establishment of 
unnatural genera such as Stygonitocrella, and (3) the blurring of 
generic boundaries. Central to this confusion stands the genus 
Nitocrella which has served as a taxonomic repository for freshwa- 
ter Ameiridae since its proposal by Chappuis (1924). Lang (1965) 
removed all species displaying 3-segmented P2-P4 endopods to a 
new genus Parapseudoleptomesochra and created a second genus, 
Pseudoleptomesochrella, to accommodate all Nitocrella species 
characterized by 2-segmented P2—P4 endopods and the presence of 
an inner seta on P2—P4 exp-1. Even under its revised taxonomic 
concept the genus Nitocrella continued to accumulate a large number 
of new species which prompted Petkovski (1976) to subdivide the 
genus even further. He suggested to group only species with 2- 
segmented P2—P4 endopods in Nitocrella s. restr. and to reallocate 
all remaining species with alternative endopodal segmentation in 
two new genera, Nitocrellopsis (P2—P3 3-segmented, P4 2-seg- 
mented) and Stygonitocrella (P2—P3 1—2-segmented, P4 
1-segmented). Petkovski (1976) also recognized three subgroups in 
Nitocrella based on the number of armature elements on P4 exp-3: 
the hirta- (3-4 setae/spines), chappuisi- (5 setae/spines) and 
vasconica-groups (6 setae/spines). Although these groups have met 
with general acceptance, their monophyletic status has never been 
challenged. Furthermore, since Petkovski (1976) did not designate a 
type species for Stygonitocrella, nor for Nitocrellopsis, both generic 
names were unavailable until recently. Galassi et al. (1999) fixed N. 
rouchi Galassi, De Laurentiis & Dole-Olivier, 1999 as the type of 
Nitocrellopsis, making the name available with their authorship. A 
similar course of action was taken by Reid ef al. (in press) who 
designated S. montana (Noodt, 1965) as the type of Stygonitocrella. 
Leg segmentation characters should be used with caution when 
inferring relationships in derived lineages. For example, the evolu- 
tionary instability of endopodal segmentation is illustrated by the 
genus Psammonitocrella Rouch which contains species with 1- (P. 
longifurcata) or 2-segmented P2—P3 endopods (P. boultoni). Unique 
derived characters such as the loss of the outer spine on P| exp-2 
leave no doubt that both Psammonitocrella species shared a com- 
mon ancestor, and hence the discrepancy observed in endopodal 
segmentation has to be interpreted as the result of intrageneric 
evolution. Utilizing endopodal segmentation patterns in defining 
generic boundaries is potentially misleading. Overweighing such 
characters at the expense of others can result in assigning species to 
47 
the wrong genus. For example, Nitocrella petkovskii Pesce, 1980 
and Stygonitocrella colchica (Borutzky & Michailova-Neikova, 
1970) are very closely related, differing essentially only in the 
expression of the segment boundary between P4 enp-1 and -2, but 
are nevertheless placed in different genera. The close zoogeographical 
connection between N. petkovskii (NW Iran) and S. colchica (W 
Georgia) is noteworthy in this context. 
I. humphreysi shows superficial similarities to both Psammoni- 
tocrella and Stygonitocrella, currently the most advanced genera 
within the Ameiridae, however the combined presence of a sexually 
dimorphic inner basal spine on Pl, a completely fused genital 
double-somite, reduced antennary exopod and vestigial PS excludes 
it from either genus. The genus Psammonitocrella was proposed by 
Rouch (1992) to accommodate two interstitial species collected in 
the hyporheic zone of an intermittent desert stream in Arizona. Its 
familial placement has been questioned by Martinez Arbizu & 
Moura (1994) who removed the genus from the Ameiridae and 
regarded it as the sistergroup of the Parastenocarididae. This course 
of action was primarily based on the loss of the outer spine of P1 
exp-2 and the absence of a sexually dimorphic inner basal spine on 
leg 1. The juvenile morphology of the P5 in both sexes and the 
presence of separate genital and first abdominal somites in the adult 
female strongly suggest a paedomorphic origin for Psammonitocrella. 
The absence of sexual dimorphism in the inner basal spine of leg 1 
should be re-evaluated in this context. The modification of this spine 
in male Ameiridae appears not until the final moult. Delaying the 
expression of this character beyond the final moult (post-displace- 
ment) would result in the secondary loss of sexual dimorphism. The 
absence of a modified inner basal spine in Psammonitocrella is 
therefore to be regarded as autapomorphic and it is proposed here to 
remove the genus from its floating status and to reallocate it to the 
Ameiridae. 
Similarities between /nermipes and Psammonitocrella are found 
in the presence of only two setation elements on the distal endopod 
segment of leg 1, the absence of the outer basal seta in P1—P2, and 
the reduced P4 endopod. It is conceivable that these shared characters 
are the product of convergence since both genera differ significantly 
in the morphology of the antenna, maxilla, maxilliped and swim- 
ming leg ornamentation. 
The supposedly cosmopolitan genus Stygonitocrella was diag- 
nosed solely on the basis of the 1-segmented P4 endopod (Petkovski, 
1976). Comparison of the swimming leg setal formulae of the 13 
species currently included strongly indicates the presence of several 
discrete lineages within this genus, each exhibiting a typical arma- 
ture pattern (Table 1) and a restricted geographical distribution 
(Table 2). This subdivision is admittedly based on swimming leg 
characters only but we suspect it to be at least partly reinforced by 
mouthpart and antennulary characters when they become available. 
Formal recognition of these lineages as distinct genera is impossible 
since most descriptions are severely lacking in detail and many of 
them are based on very few specimens or one sex only. For example, 
the description of P. djirgalanica is completely lacking in illustra- 
tions (Borutzky, 1978). In addition, the type material of the great 
majority of its species is no longer extant and additional records 
have not been added. S. petkovskii differs from its congeners in the 
absence of the inner seta on P3 exp-2, an element which is present in 
all other congeners. Attempts to trace the single female on which this 
description was based failed (Galassi, pers. comm. ). 
As pointed out by Reid er al. (in press) the generic placement of 
S. orghidani (Petkovski, 1973) is questionable. The original descrip- 
tion is very concise, showing illustrations only of the antenna, 
caudal rami and the fifth legs. The exopodal armature of the swim- 
ming legs is largely unknown, apart from Petkovski’s statement that 
