Biological Journal of the Linnean Society, 2007, 92, 367^01. With 16 figures 
Ancestry to an endemic radiation in Lake Tanganyika? 
Evolution of the viviparous gastropod Potadomoides 
Leloup, 1953 in the Congo River system 
(Caenogastropoda, Cerithioidea, Paludomidae) 
MATTHIAS GLAUBRECHTi* and ELLEN E. STRONG^ 
^Museum fiir Naturkunde, Humholdt University, Department of Malacozoology, Invalidenstrajie 43, 
D-10115 Berlin, Germany 
^Smithsonian Institution, National Museum of Natural History, PO Box 37012, MRC 163, 
Washington, DC, 20013-7012, USA 
Received 8 June 2006; accepted for publication 20 December 2006 
Providing another spectacular model for understanding speciation and radiation, the origin of the gastropod species 
flock in Lake Tanganyika (with an estimated age of approximately 12 Myr) remained enigmatic to date. Although, 
for a long time, an in situ radiation was assumed, Lake Tanganyika could have functioned as a reservoir for ancient 
African lineages, implying that the now lacustrine taxa originiated elsewhere. However, the fluviatile gastropod 
fauna of adjacent river systems in Central and East Africa is only poorly known. Here, we provide conchological, 
anatomical, phylogenetical, and biogeographical data on the fluviatile genus Potadomoides Leloup, 1953, which was 
hitherto regarded as ancestral to the entire Tanganyika gastropod radiation. The type species Potadomoides 
pelseneeri is restricted to the delta region of the Malagarasi River east of Lake Tanganyika, whereas three congeneric 
species (Potadomoides bequaerti, Potadomoides hirta, and Potadomoides schoutedeni) inhabit the Congo River with 
its tributaries Lualaba and Luvua, west of the Tanganyikan Rift. We describe and document, with scanning electron 
microscopy, the ontogenetic development of embryos of this uterine brooder as well as the detailed reproductive 
anatomy. Phylogenetic analysis of 44 morphological characters (including adult and embryonic shell, operculum, 
radula, reproductive tract) for 15 paludomid taxa could not support monophyly of the Tanganyika species flock. 
Instead, we found two major lineages that colonized Lake Tanganyika independently, one comprising the 
Nassopsinae Kesteven, 1903 (= Lavigeriinae Thiele, 1925) with the riverine Potadomoides plus the lacustrine 
Lavigeria and Vinundu, the second comprising the riverine Cleopatra together with the rest of the lacustrine species 
(except for Tiphobia horei). The analysis identifles Potadomoides as paraphyletic, with the uterine brooder 
P. pelseneeri being the sister taxon to the uterine brooder Lavigeria plus the oviparous Vinundu, but not to the entire 
Tanganyika species flock. We reconstruct the independent evolution of an fluviolacustrine taxon Nassopsinae for 
which we evaluate the synapomorphic characters, in particular those of reproductive biology, and discuss systematic 
and evolutionary implications of repeated origin of (ovo-)viviparity in these limnic Cerithioidea. Finally, we outline 
a hypothesis on the evolutionary history of Potadomoides in the context of the gastropod radiation in Lake 
Tanganyika.    ? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01. 
ADDITIONAL KEYWORDS: evolutionary ecology ? fluviatile fauna ? species flock ? thalassoid gastropods ? 
uterine brooding ? viviparity. 
INTRODUCTION regarded as hotspots of aquatic biodiversity and as 
natural laboratories  providing insights  into  evolu- 
Ancient  lakes  with  their  endemic   species   assem- ^:^^^^^ processes, such as intralacustrine speciation 
blages,  like  Lake  Tanganyika  in  East Africa,   are (Brooks,    1950;    Boss,    1978;    Fryer,    1991,    1996; 
 Martens, Coulter & Goddeeris, 1994; Martens, 1997; 
*Corresponding author. Rossiter & Kawanabe, 2000). This is exemplified by 
E-mail: matthias.glaubrecht@museum.hu-berlin.de recent  studies  on the  origin  of vertebrate  species 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367-401 367 
368     M. GLAUBRECHT and E. E. STRONG 
flocks particularly in cichlids (Salzburger et al., 2002; 
Streelman & Danley, 2003; Verheyen etal, 2003; 
Kocher, 2004, 2005; Salzburger & Meyer, 2004), but 
now also in invertebrates such as gastropods 
(Rintelen & Glaubrecht, 2003, 2005; Rintelen etal, 
2004; Wilson, Glaubrecht & Meyer, 2004). A long- 
standing assumption has been the antiquity of 
ancient lake lineages and, closely correlated, a 
usually highly derived morphology (Brooks, 1950; 
Nishida, 1991; Martens, 1997). Although current 
research has indicated the potential for rather rapid 
speciation as well as morphological and genetic diver- 
sification in organisms prone to sexual selection as, 
for example, cichlids (Sturmbauer, 1998; Danley & 
Kocher, 2001; Streelman & Danley 2003), recent 
work has found evidence for an ancient origin of 
disparity and diversity even predating lake formation 
of Tanganyika in thalassoid (i.e. 'marine-like') gastro- 
pods (Wilson et al., 2004) and cichlids (Joyce et al., 
2005). This alternative interpretation of lacustrine 
radiations as surviving remnants and the lake as an 
evolutionary reservoir, as hypothesized by Banister & 
Clarke (1980) and Nishida (1991), contradicts the 
long and widely held belief of in situ radiations 
(Brooks, 1950; Boss, 1978; Coulter, 1991; Michel et al, 
1992; Michel, 1994, 2000). Focusing on the lacustrine 
fauna alone overlooked the potential nonlacustrine 
origin of many of these lineages, as recently sug- 
gested for cichlid fishes (Joyce et al., 2005). 
Despite the considerable interest in the evolution- 
ary biology of the fauna in East African lakes and 
their endemic species flocks (Coulter, 1991; Fryer, 
1991, 1996; Martens, 1997; Rossiter & Kawanabe, 
2000), the origin, systematics and phylogenetic 
relationships of the lacustrine gastropods of Lake 
Tanganyika remained largely unresolved to date. The 
ancestry of the riverine Potadomoides Leloup, 1953, 
to thalassoid gastropods in the lake has long been 
assumed but never tested (Brown, 1994: 129, 530; see 
also Brown & Mandahl-Barth, 1987: 318; Coulter, 
1991: 237), although morphological evidence evalu- 
ated by Glaubrecht (1996: 148-150) hinted at a close 
relationship with the lacustrine Lavigeria from Lake 
Tanganyika. Recent studies on the phylogeny of fresh- 
water Cerithioidea in general, and in particular of 
the formerly so-called African 'thiarids' (Glaubrecht, 
1996, 1999, 2006; M. Glaubrecht, unpubl. data; 
Lydeard etal, 2002; Wilson etal, 2004), revealed 
new insight also into the classification of the Tanga- 
nyikan taxa. Accordingly, we here treat the thalassoid 
gastropod fauna of Lake Tanganjdka and Potado- 
moides as members of the family Paludomidae. 
The first collections of Potadomoides were made at 
the upper Congo River by Joseph Bequaert in 1910, 
and described by Dautzenberg & Germain (1914) who 
allocated   several   species   to   the   genus   Cleopatra 
Troschel, 1857; this placement was followed later by 
Pilsbry & Bequaert (1927: 296-298). The marine-like 
appearance of these Congo species was explicitly 
pointed out when Dautzenberg & Germain (1914: 2) 
noted that 'ces mollusques ont un aspect halolimnique 
indeniable'. Later, in a first comprehensive review on 
ancient lakes. Brooks (1950: 150) stated that 'the 
thalassoid snails [of Lake Tanganyika] presumably 
are derived from fluviatile ancestors which were 
present at the lake's genesis, but these descendants 
have so far given no clue to the relationships of their 
ancestors'. Mandahl-Barth (1967) re-studied riverine 
gastropods from the Congo system and described the 
radulae of some of these so-called Cleopatra taxa, 
transferring five species to Potadomoides that had 
been named by Leloup (1953) to accommodate the 
type species Potadom,oides pelseneeri from the delta of 
the Malagarasi River flowing into Lake Tanganyika 
from the east. 
The present study aims to provide and evaluate 
new data on the morphology and conchology using 
histology and scanning electron microscopy (SEM) 
(including embryonic features of shells, radula, repro- 
ductive biology, and ontogeny) as well as the biogeo- 
graphy of the constituent species of Potadomoides in 
the Congo and Malagarasi River. In addition, using 
comparative morphological data from an ongoing 
study of the thalassoid species flock in Lake Tanga- 
nyika, we analyse the phylogenetic systematics of this 
fluviatile taxon in relation to the endemic lacustrine 
gastropods. The phylogeny generated here allows us 
to re-evaluate the long-held hypothesis with respect 
to the ancestry of this radiation and to address 
general evolutionary questions with respect to to the 
historical process of species flock formation in this 
ancient lake. We discuss the evolution of uterine 
brooding within a fluviolacustrine taxon Nassopsinae 
Kesteven, 1903 (= Lavigeriinae Thiele, 1925; for 
nomenclatural details, see Bouchet & Rocroi, 2005). 
Based on data from palaeogeography and palaeohy- 
drology, we present a hypothesis on the evolutionary 
history of Potadomoides in the Congo River system in 
the context of the lacustrine speciation of the thalas- 
soid gastropods in Lake Tanganyika. 
MATERIAL AND METHODS 
MATERIAL 
All specimens of Potadom,oides species obtainable 
from museum collections were studied, thus essen- 
tially comprising the material in two Belgian institu- 
tions, the Musee Royal L'Afrique Centrale (Tervuren) 
and the Institut Royal des Sciences Naturelles de 
Belgique (Bruxelles), with additional material located 
in the Museum of Comparative Zoology, Harvard Uni- 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     369 
versity, Cambridge, MA, USA and the Danish Bilhar- 
ziasis Laboratory, Charlottenlund, Denmark. We are 
not aware of any more recent collections of relevant 
material from the Congo River. 
'right' describe features of the midgut as opened 
along a dorsal (exposed surface) longitudinal incision 
with the roof deflected to the left and the style sac 
uppermost. 
ANATOMICAL INVESTIGATIONS 
Radulae for P. pelseneeri and Potadomoides schout- 
edeni were prepared from alcohol preserved material 
and re-hydrated soft bodies found within the shells, 
respectively. In addition, in the Danish Bilharziasis 
Laboratory, we located radulae for Potadomoides 
hequaerti, P. schoutedeni and Potadomoides hirta 
mounted in Canada Balsam by Mandahl-Barth 
(1967). Because these were the only existing radula 
preparations for P. bequaerti and P. hirta, the radulae 
for these species were examined and photographed 
under a compound microscope, allowing us to verify 
the congeneric identity of two Potadom,oides species 
otherwise only known from inconclusive drawings. 
Formula for radular tooth descriptions: (1) rachid- 
ian: numbers of denticles on the left side/median 
denticle/numbers of denticles on the right side; (2) 
lateral teeth: inner cusps/pronounced denticle/outer 
cusps; (3) marginal teeth: number of cusps on inner 
marginal tooth + number of cusps on outer marginal 
tooth. 
For SEM examination of P. pelseneeri, embryos 
were removed from the uterine brood pouch, treated 
with hexamethyldisilizane following the procedure 
previously described by Nation (1983) and photo- 
graphed with a Jeol 6300F Scanning Microscope at 
the Zoologisches Museum Berlin. Due to the abbre- 
viation or loss of early ontogenetic stages in vivipa- 
rous freshwater Cerithioidea, a distinct sculptural 
transition from the embryonic (primary) shell and 
larval (secondary) shell to the adult or tertiary shell 
(= teleoconch) is lacking. Therefore, the term 'proto- 
conch' as applied to the first two stages in oviparous 
forms is not comparable with the conditions found 
among viviparous gastropods and, consequently, we 
prefer the term 'embryonic shell' for all shelled stages 
in the brood pouch. Methods and terminology for 
embryonic shell measurements follow Glaubrecht 
(1996). Height (h), width (w) and maximum diameter 
at one whorl (d) were obtained from scanning electron 
micrographs. For shell microstructure, adult shells 
were broken perpendicular to the outer surface of the 
shell and photographed with a Jeol 6300F Scanning 
Microscope at the Natural History Museum Berlin. 
Anatomical dissections were performed using a 
Leica MZ 9.5 binocular microscope with camera 
lucida. For histology the pallial oviduct of J! bequaerti 
was embedded in paraplast, sectioned at 6 \jja\ and 
stained with haematoxylin and eosin. For midgut 
descriptions, topological relations such as 'left' and 
PHYLOGENETIC ANALYSIS 
Phylogenetic analysis was performed using 44 mor- 
phological characters, particularly using features 
from adult and embryonic shell, operculum, radula, 
and reproductive tract described herein for Potado- 
moides, supplemented by published accounts for 
other taxa (Strong & Glaubrecht, 2002, 2003, 2007; 
Glaubrecht & Strong, 2004; Michel, 2004); if not 
stated otherwise, we used additional data from our 
own anatomical studies (E. E. Strong & M. Glaubre- 
cht, unpubl. data). 
Editing of the data matrix was completed using 
WinClada (Nixon, 1999-2002) (Table 3). Phylogenetic 
analyses were performed with the Ratchet as imple- 
mented in WinClada using default options (200 itera- 
tions); a more rigorous search did not yield any 
further topologies. Characters were analysed as unor- 
dered. The distribution of the characters on the con- 
sensus tree (Fig. 15) was examined using the fast 
optimization option implemented in WinClada. 
Bootstrap support was also calculated in WinClada 
with the following parameters set: 1.000 replications, 
100 search replicates per replication and ten starting 
trees per replicate. 
According to cladistic analyses of Cerithioidea, in 
particular those limnic taxa formerly subsumed 
in 'Thiaridae' (Glaubrecht, 1996, 1999, 2006; M. 
Glaubrecht, E. E. Strong, J. Healy & W. Ponder, 
unpubl. data), we chose to root the tree with Mel- 
anoides tuberculata, a procedure consistent with 
other analyses (Michel, 2000; West & Michel, 2000; 
Wilson etal., 2004). A total of 15 ingroup taxa was 
included: in addition to the four Potadomoides species 
and Lavigeria and Vinundu, we also included five 
other thalassoid taxa to represent different lineages 
within the species fiock as recovered in former analy- 
ses (West & Michel, 2000; Michel, 2004; Wilson et al., 
2004), and three African species of Cleopatra plus the 
Asian Paludomus. 
CODENS 
The name used throughout the present study for 
the former Belgian Congo, later known as Zaire 
(until 1997), is Democratic Republic of Congo (DRC). 
Institution codens are: DBL, Institute of Health 
Research (formerly Danish Bilharziasis Laboratory, 
Charlottenlund (Denmark); IRSNB, Institut Royal 
des Sciences Naturelles de Belgique, Bruxelles 
(Belgique); MCZ, Museum of Comparative Zoology, 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
370     M. GLAUBRECHT and E. E. STRONG 
Harvard University, Cambridge, MA (USA); MRAC, 
Musee Royal L'Afrique Centrale, Tervuren (Bel- 
gique); ZMB, Museum fiir Naturkunde, Humboldt 
University (formerly Zoologisches Museum Berlin). 
SYSTEMATICS 
PALUDOMIDAE 
POTADOMOIDES LELOUP, 1953 
Cleopatra Troschel, 1857 - Dautzenberg & Germain 
(1914) 
Potadomoides   Leloup,   1953:   102 - Mandahl-Barth, 
1967: 127; Brown & Mandahl-Barth, 1987: 308, 318; 
Brown, 1994: 129 
Taxonomic remarks 
So named for the superficial similarity of the shell 
and operculum to those of Potadoma Swainson, 1840, 
although the radula is different from this genus 
where it is typically pachychilid-like. Potadomoides 
differs from Cleopatra in its paucispiral operculum 
and the fewer number of cusps on the central tooth. 
Type species, by monotj^y, is Potadom,oides pelse- 
neeri Leloup, 1953. Four additional species previously 
assigned to Cleopatra, by Dautzenberg & Germain 
(1914) were allocated later to the genus by Mandahl- 
Barth (1967). Congenerity is herein verified for three 
of these additional species, namely P. bequaerti, hirta 
and schoutedeni, whereas broecki from the Aruwimi 
River is suggested to be better assigned to Cleopatra 
instead, based on conchological and radular features 
(M. Glaubrecht, unpubl. data). 
Diagnosis 
Shell small to medium-sized, but solid, ovate with 
convex to pronouncedly angulated whorls, sculptured 
in various ways (i.e. with only low spiral ridges or 
elevated carinae and spines). Embryonic shell with 
wrinkled sculpture on first whorl, followed by pro- 
nounced axial elements crossing spiral lines, thus 
creating a reticulate sculpture. Operculum comeus, 
paucispiral, with subcentral nucleus on columellar 
side. Mantle border smooth. Radula relatively strong, 
with small, narrow rachidian; cutting edge with one 
median, rounded cusp and up to three, but mostly one 
or two smaller flanking denticles; basal rim V-shaped. 
Upward curved, hook-shaped cusp at upper lateral 
side of rachidian. Laterals with major prominent, 
spatulate cusp, mostly one inner denticle and rounded 
cusp flanked on outer side by up to two or three 
denticles; with short, wing-shaped lateral extensions, 
and well developed basal tongue. Marginals dissimilar, 
inner marginals with mostly one (sometimes a second 
smaller) inner denticle, a very large broadly rounded, 
shovel-shaped main cusp, flanked on the outside by 
one or two small cusps; the second, outer marginal 
often with ten finger-shaped cusps (but only eight in 
the type species). Gonoducts closed; pallial oviduct 
(= uterus) functioning as brood sac, filled with embryos 
(P. bequaerti) and/or shelled juveniles (P. pelseneeri). 
Distribution and ecology 
Endemic to tributaries of the Congo River system 
in DRC (three species) and the Malagarasi River 
delta on the eastern shore of Lake Tanganyika 
(P. pelseneeri). Although P. bequaerti, P. hirta and 
P. schoutedeni appear to be rheophilic, P. pelseneeri is 
fluviatile with an occurrence apparently restricted 
to the delta of the Malagarasi River (Fig. 1). 
Fossils 
Van Damme (1984) and Van Damme & Pickford 
(2003) reported that several fossil species referred to 
the genus Platymelania with an appearance resem- 
bling Potadomoides are known from early Pleistocene 
(Villafranchian) deposits in the Lake Albert?Edward 
Rift system, in particular the area of the Congo and 
Uganda sides (Gautier, 1970). Several species are 
known, such as, for example, Platymelania brevis- 
sima (Cox, 1926) from the Kaiso beds of Lake Albert 
with strong, oblique ribs (Gautier, 1970: 110 flf.), 
or Platymelania bifidicincta (Cox, 1926) from the 
Kazinga Channel, Lake Edward beds, and Sinda beds 
at the Semliki-River connecting to Lake Albert, char- 
acterized by a bifurcating keel on the last whorl as 
diagnostic feature. For a recent treatment and phy- 
logenetic interpretation of fossils in the absence of 
rigorous cladistic analyses, see Van Damme & 
Pickford (2003). It should be noted that the specimens 
figured as fossil Platymelania and discussed in 
context of the Recent Potadomoides often resemble 
fossil Cleopatra or Viviparus, to which they have 
also been assigned by other authors. Thus, their 
affinities to lacustrine or riverine paludomids remain 
uncertain. 
Remarks 
Judging from their absence in almost all malacologi- 
cal collections, species of Potadomoides appear to be 
fairly rare and restricted in their occurrence. To date, 
they are known primarily from the type material, 
thus those animals collected by the only expedition 
to the Congo River by Joseph Bequaert in 1910 
(Dautzenberg & Germain, 1914), and by Leloup in 
1946?47 with the Belgian Exploration Hydrob- 
iologique of Lake Tanganika (Leloup, 1953). The 
primary and secondary types in the IRSNB and 
MRAC are supplemented by but a few other second- 
ary types distributed, for example, to the MCZ and 
DBL. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     371 
Potadomoides 
^    ^^^   Tanganyika 
Figure 1. Disjunct distribution of species of Potadomoides Leloup, 1953 in the Congo River drainage, with tributaries 
predating the formation of Lake Tanganyika. *, Potadomoides pelseneeri, the type species in the Malagarasi River delta; 
A, Potadom,oides bequaerti; M, Potadom,oides hirta; ?, Potadom,oides schoutedeni. Note that a fifth species, broecki which 
is known to occur in the Aruwimi River of the northern Oriental Province of the Democratic Republic of Congo, is 
suggested to be re-transferred to Cleopatra (see text for details). Shell figures not to scale. Scale bar = 100 km. 
POTADOMOIDES PELSENEERI LELOUP, 1953 
Potadomoides pelseneeri - Leloup, 1953: 102?105, fig- 
ures 47, 57p, 72h, pL 3, figure 6; Mandahl-Barth, 
1967: 128; Brown & Mandahl-Barth, 1987: 308, 318; 
Brown, 1994: 129, figure 62b; Glaubrecht, 1996: 148- 
150, pi. 21, figures 5, 6, 7, 8, 9. 
Type locality 
Tanzania: 'Dans le delta de la Malagarasi'; on the 
eastern shore of Lake Tanganyika, south of Ujiji 
(Fig. 1). 
Type material 
Holotype and 72 paratypes (IRSNB no. 145). The 
holotype (non vidi) represents a shell containing the 
soft body still inside (C. Massin, pers. comm.). One 
paratype comprises  a female preserved in ethanol 
with the soft part isolated from the shell and with 
juveniles in the uterus (specimen studied and photo- 
graphed in 1993 by M. Glaubrecht; Fig. 9). Unfortu- 
nately, this paratype is currently untraceable due to 
misplacement in the IRSNB collection (C. Massin, 
pers. comm.). Two other paratypes comprising shells 
with soft bodies were examined. Of these, one shell 
was cracked, the soft body removed, and used for 
dissection and preparation of embryos for SEM; a 
second specimen represented a juvenile in poor con- 
dition and was used for radula preparation (Fig. 6D, 
F). The remaining paratypes (AT =69) are empty 
shells lacking preserved soft parts, of which four are 
depicted in Figure 2A. Note that we here use the holo- 
and paratype assignments indicated by the IRSNB to 
prevent   further   confusion.   However,   the   original 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
372     M. GLAUBRECHT and E. E. STRONG 
Figure 2. Shells of Potadomoides. A, Potadomoides pelseneeri, four paratypes (IRSNB, no. 145). B, Potadomoides 
bequaerti, from left to right: lectotype (MRAC 47.358) and three paralectotypes (MRAC 47.358a, 47.359, 47.351); note 
variation in shell sculpture ranging from carinated to rounded whorls. C, Potadomoides hirta, two syntypes (MRAC 
47.432-33). D, Potadomoides schoutedeni, lectotype (MRAC 47.804) and four paralectotypes (MRAC 47.805, 47.812-14). 
Scale bar = 5 mm. 
description by Leloup (1953) does not explicitly des-       Other matenal examined 
ignate  holo-   and/or  paratypes   among  his   original       No other samples than those present in the IRSNB 
series, rendering them all syntypes. collection were available to us. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     373 
Figure 3. Embryonic shells of Potadomoides pelseneeri (A?F) and of Potadomoides schoutedeni (G?I). (A, B) Apertural 
view of embryonic shell (IRSNB, no. 145) at approxomately one whorl. Scale bar = 100 ^m. B, apertural view at 
approximately 1.5 whorls. Scale bar = 100 |xm. C, B, D, Apical view of embryonic shell; note individual variation. Scale 
bar = 100 |xm. E, Detail of apical view; note irregular texture and transition zone to accretionary growth. Scale 
bar = 30 |xm. F, shell microstructure with three crossed lamellar layers; outer shell surface is up, arrows indicate 
transitions between layers. Scale bar = 500 |xm. G, H, I, embryonic shell of i-1 schoutedeni preserved in a juvenile specimen 
(MRAC 47.807?11). G, apertural view of juvenile shell. Scale bar = 1000 |xm. H, apical view of embryonic shell. Scale 
bar = 1000 |xm. I, detail of apical view; note that outer layers are partly eroded. Scale bar = 30 |xm. 
Taxonomic remarks 
The species is named after P. Pelseneer, one of the 
first authors to discuss the hypothesis of a marine 
origin for Lake Tanganyika's fauna and the causes of 
the thalassoid appearance of its endemic molluscs 
(Pelseneer, 1886: 112-115). 
Description 
Shell (Fig. 2A); Ovate, thick-walled with five to six 
slightly convex whorls; lip thickened, white, holo- 
stomous; last whorl weakly carinated at periphery, 
with fine spiral ridges, strongest at base. Last whorl 
largest, comprising approximately two-thirds of shell 
height; aperture measuring approximately two-fifths 
of the total shell height. Of the paratypes (IRSNB 
no. 145), 55 shells measured have the following 
average parameters (mean ? SD): height = 7.19 ? 
1.86 mm; width = 4.72 ? 0.99 mm; height/width 
index = 66.90 ? 6.30 (Table 1). Specimens range in 
size from 3.0x2.2 to 10.6 x 6.7 mm (Fig. 4). Shell 
comprising four to five layers, namely a thin outer 
irregular prismatic layer and a thin inner regular, 
simple prismatic layer, bounding two crossed lamellar 
layers in early whorls, and an additional crossed 
lamellar layer in later whorls (Fig. 3F). 
Embryonic shell (Fig. 3A?E).- Sculpture of first whorl 
coarsely pustulose and pitted; apical tip of embryonic 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
374     M. GLAUBRECHT and E. E. STRONG 
14.0 n 
12.0 - 
E     8.0 - 
E 
4.0 - 
2.0 - 
0.0 
o o 
o <i^o 
J?^^ *" 
? m? 
A o 
Oo 
4.0 5.0 
width (mm) 
Figure 4. Shell shape given as height versus width of Potadomoides species; O, Potadomoides pelseneeri; ?, Potado- 
moides bequaerti; A, Potadomoides schoutedeni; ^ = Potadomoides hirta; M, Cleopatra broecki is added for comparison 
(see text for detailed explanation). Note that both P. hirta and Cleopatra broecki generally have broader shells than (other) 
Potadomoides species. 
shell slightly inflated, overl3dng subsequent whorl. 
Coarse, irregular (i.e. wrinkled) ornament diminish- 
ing following transition from flrst to second whorl, 
rapidly replaced by numerous flne, undulating spiral 
ridges and weak growth lines, creating a character- 
istic reticulate pattern. 
Operculum (Fig. 5A, op); Ovate, corneus, translucent, 
thinning to outer edges, light amber brown in colour; 
paucispiral with subcentral nucleus on columellar 
side. 
External anatomy: Anterior pedal gland opening to 
long, shallow slit along propodium. Snout large, tri- 
angular, dorso-ventrally flattened. Retracted cephalic 
tentacles approximately equal in length to snout. 
Small eyes located half way from base of retracted 
tentacles on slight protuberances (Fig. 5B). Mantle 
edge smooth. Ctenidium with large triangular fila- 
ments, extending from base of pallial cavity to mantle 
margin (Fig. 5B, ct). Osphradium (os) forming 
smooth, narrow ridge in shallow depression along 
ctenidial axis bounded by low flap. Hypobranchial 
gland weakly developed. 
Radula (Fig. 6A-F): Relatively strong, with 52 rows 
in a juvenile and approximately 70 rows in the 
radula of the adult studied; the latter being in poor 
condition and thus slightly incomplete in the radula 
sac portion. Rachidian teeth narrow, slightly more 
than 1.5-fold longer than wide, tapering to V-shaped 
base. Cutting edge bearing 1/1/1 prominent, smooth 
denticles. Central denticle large, bluntly rounded, 
with flanking accessory denticles smaller and rather 
pointed. Upper lateral edge with hook-like 'denticle', 
curving pronouncedly up in some rows. Lateral teeth 
bearing 1(2)/1/1(2) denticles; central denticle large, 
spatulate and bluntly rounded; flanking denticle(s) 
also rounded and approximately one-half the length 
of the central denticle. Inner lateral edge straight, 
tapering along lower third to prominent V-shaped 
basal, pointed projection. Lateral extensions short, 
approximately equal in length to cusped portion of 
tooth. Marginal teeth moderately long, with flat- 
tened shafts tapering to narrow, pointed basal edge. 
Inner and outer marginal teeth dissimilar. Inner 
marginal teeth with broad shafts bearing large 
shovel-shaped, bluntly rounded main denticle 
flanked by one broad inner denticles and one 
smaller, more pointed outer denticle. Outer marginal 
teeth with narrow shafts bearing eight rounded, 
flnger-like denticles of similar shape and size. 
Radula of one juvenile studied (Fig. 6D, E) with nar- 
rower rachidian tooth (1/1/1). Denticle pattern of 
most teeth quite similar to that of the adult speci- 
men, although laterals appear to have more elon- 
gate, pointed cusps and 1/1/2?3 pattern, and 
sometimes nine cusps in the outer marginals. 
Foregut: Mouth opening at ventral tip of snout. 
Buccal mass robust. Large odonotophore occupying 
posterior two-thirds of buccal cavity. Small, rounded. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     375 
Figure 5. Anatomy of Potadomoides pelseneeri. A, external anatomy, dorsal view (dotted line indicates extent of 
pericardial cavity under kidney; stippling of kidney represents excretory tubules). B, anatomy of cephalic haemocoel, 
dorsal view. C, kidney morphology, dorsal view. Internal features of bladder and pallial kidney extension revealed by 
partially removing intestine. Dotted line represents small chamber behind afferent renal vessel between kidney and 
pericardial wall. Arrow indicates communication between bladder and main kidney chamber. D, circum-oesophageal nerve 
ring, posterior view on left, lateral view on right. Scale bar = 1 mm. arv, afferent renal vessel; bg, buccal ganglion; bp, 
brood pouch; ce, cerebral ganglion; cm, columellar muscle; ct, ctenidium; dgl, digestive gland; e, oesophagus; go, gonad; 
int, intestine; kDa, kidney; np, nephropore; op, operculum; os, osphradium; pd, pedal ganglion; pi, pleural ganglion; per, 
pericardium; ret, buccal mass retractor; rs, radular sac; s, septum; sc, statocyst; sb, suboesophageal ganglion; sgd, salivary 
gland duct; sp., supra-oesophageal ganglion; st, stomach. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
376     M. GLAUBRECHT and E. E. STRONG 
Figure 6. Radula of Potadomoides; scanning electron micrographs. A, B, C, Potadomoides pelseneeri, adult (paratype, 
IRSNB no. 145). D, E, F, Potadomoides pelseneeri, juvenile (paratype, IRSNB no. 145). G, H, I, Potadomoides schoutedeni 
(paratype, MRAC 47.812-814). A, Anterior radular ribbon. Scale bar = 100 |xm. B, rachidian and lateral teeth. Note 
V-shaped base of rachidian. Scale bar = 30 |xm. C, rachidian, lateral and marginal teeth. Scale bar = 30 |xm. D, anterior 
radular ribbon. Scale bar = 10 ^im. E, rachidian and lateral teeth. Scale bar = 10 ^im. F, lateral and marginal teeth. Scale 
bar = 10 |xm. G, anterior radular ribbon. Scale bar = 30 |xm. H, rachidian and lateral teeth. Scale bar = 10 ^im. I, lateral 
and marginal teeth. Note V-shaped base of rachidian. Scale bar = 30 |xm. 
glandular subradular organ lying at anterior end of 
odontophore in shallow sublingual cavity. Deep, glan- 
dular, complexly folded buccal pouches (Fig. 5B, bpo) 
extending posteriorly from buccal ganglia (bg) at 
posterior end of buccal cavity into anterior oesopha- 
gus; right pouch longer, extending farther anteriorly 
into buccal cavity. Pouches contiguous along ventral 
midline of anterior oesophagus, separated anteriorly 
by smooth, triangular field of epithelium in posterior 
buccal cavity overlying radula; each pouch bounded 
by thin, longitudinal fold. Jaws weakly developed at 
anterior limit of dorsal folds. Salivary glands (sgd) 
forming long, unbranched tubules, extending through 
nerve ring, opening dorso-laterally alongside odonto- 
phore. Paired buccal retractors (ret) inserting on 
lateral walls of cephalic haemocoel; retractors placed 
asymmetrically, right retractor inserting anteriorly 
near base of cephalic tentacle, left retractor inserting 
posteriorly just in front of circum-oesophageal nerve 
ring.  Radular sac (rs)  long,  curving around right. 
posterior end of buccal mass. Anterior oesophagus 
bearing paired, grooved longitudinal folds extending 
posteriorly from tips of buccal pouches. Mid- 
oesophageal gland lacking (e). 
Midgut: Oesophagus opening ventrally on left side of 
midgut (Fig. 7, e). Marginal fold (mf) passing ante- 
riorly from oesophageal aperture, then turning 
posteriorly bordering right edge of sorting area (sa). 
Sorting area triangular, tapering posteriorly, with 
straight left margin. Accessory marginal fold (amf) 
emerging from left side of oesophageal aperture, par- 
alleling marginal fold around posterior tip of sorting 
area, bifurcating for short distance to form two folds. 
Midgut roof alongside sorting area cuticularized (cu); 
epithelium transversely folded anteriorly and forming 
elongate, rectangular, smooth pad posteriorly. Large 
gastric shield (gs) continuous with cuticle of stomach 
roof and crystalline style pocket (p). Glandular pad 
(gp) large, rounded with small, rounded protuberance 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     377 
amf 
Figure 7. Midgut morphology oiPotadomoides pelseneeri, dorsal view. Midgut opened laterally on the right, roof reflected 
to the left; anterior is uppermost (coarse stippling indicates cuticle lining roof). Scale bar = 1 mm. amf, accessory marginal 
fold; ap, accessory pad; c, caecum; cr, crescentic ridge; cu, cuticle lining stomach roof; e, oesophagus; gp, glandular pad; 
gs, gastric shield; mf, marginal fold; p, crystalline style pocket; sa, sorting area; ss, style sac lip; tl, major typhlosole; 
uf, 'u-shaped' fold. 
at anterior end to left of gastric shield. Crescentic 
ridge (cr), bounding deep crescentic groove, extending 
back from oesophagus and fusing to right side of 
glandular pad. Paired digestive gland ducts opening 
to deep pocket between proximal end of crescentic 
ridge and glandular pad. Shallow caecum (c) extend- 
ing ventrally under glandular pad behind gastric 
shield. Single, fine, longitudinal fold along floor 
behind gastric shield, opposite caecum. Field of fine 
parallel folds to right of marginal fold extending ante- 
riorly from oesophagus to base of major typhlosole 
(tl). Prominent fold (uf) along left side of cuticular- 
ized crystalline style pocket extending anteriorly to 
base of pocket below style sac lip (ss), bounding 
U-shaped depression. Style sac communicating with 
intestinal groove along proximal one-fourth of style 
sac. Crystalline style present. 
Hindgut: Proximal intestine (Fig. 5A, int), lying 
alongside style sac, passing under distal end of style 
sac behind pericardium. Intestine extending posteri- 
orly alongside style sac to anterior end of gastric 
chamber (st), curving under posterior tip of kidney 
(kd), and continuing forward along pallial gonoduct 
(bp) to papillate anus near mantle margin. 
Reno-pericardial system: Kidney (Fig. 5A, kd) long, 
narrow, lying obliquely across body axis. Kidney 
bounded posteriorly by digestive gland (dgl), 
anteriorly overhanging base of mantle cavity and 
extending into pallial cavity alongside gonoduct, 
ventral to intestine (Fig. 5C, int). Lumen partially 
subdivided into several chambers. Main chamber 
containing large mass of excretory lamellae (stippled 
region) dorsally overlying voluminous bladder 
(exposed chamber) alongside pericardium (per); 
bladder communicating dorsally to main chamber 
above via small aperture (arrow) short distance 
behind nephropore (np). Bladder posteriorly par- 
tially subdivided by incomplete horizontal septum 
(s) extending from nephropore between intestine 
and common reno-pericardial wall. Bladder extend- 
ing anteriorly into pallial roof below intestine, 
alongside gonoduct; pallial extension separated into 
right and left, laterally compressed chambers, com- 
municating posteriorly near nephropore. Several 
small branches of afferent renal vessel (arv) supply- 
ing right and left pallial chambers and ventral floor 
of bladder. Afferent renal vessel separating small 
chamber (dotted line) between kidney and pericar- 
dium.   Nephridial  gland  absent.   Pericardial  cavity 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
378     M. GLAUBRECHT and E. E. STRONG 
Figure 8. Reproductive anatomy otPotadomoides pelseneeri. A, left lateral view, with anterior to the left. B, right lateral 
view, with anterior to the right. C, internal view. Scale bar = 1 mm. agl, albumen gland; cgl, capsule gland; emb, embryo; 
ovi, renal oviduct; res, seminal receptacle; spb, spermatophore bursa; vc, ventral channel. 
long, deep, narrow, extending underneath kidney 
alongside style sac to posterior intestinal loop. 
Auricle communicating with gill via efferent bran- 
chial vein and with kidney. 
Nervous system: Circum-oesophageal nerve ring lying 
just behind base of buccal mass. Rather elongate 
commissure connecting cerebral ganglia (Fig. 5D, ce), 
each producing five stout nerves. Pleural ganglia (pi) 
lying behind and below cerebral ganglia, connected to 
cerebral ganglia via short connectives. Pedal ganglia 
(pd) with three prominent anterior pedal nerves and 
three to four smaller accessory nerves. Statocysts (sc) 
with numerous statoconia present on posterior, dorsal 
surfaces of pedal ganglia. Sub-oesophageal ganglion 
(sb) connected to left pleural ganglion via short con- 
nective. Right dialyneury present between nerves of 
suboesophageal and right pleural ganglia. Long con- 
nective joining right pleural and supra-oesophageal 
ganglia (sp.). Left dialyneury formed between pallial 
nerve of left pleural ganglion and osphradial nerve of 
supra-oesophageal ganglion near junction of pallial 
roof and floor. 
Reproductive biology: Gonad (Fig. 5A, go) dorsally 
overlying digestive gland from tip of visceral mass to 
posterior end of gastric chamber (st). Renal oviduct 
(Fig. 8B, ovi) forming short, recurvent loop before 
entering base of glandular pallial oviduct. Pallial 
oviduct, with proximal albumen (Fig. 8B, C, agl) and 
distal capsule (cgl) glands, opening to mantle cavity 
via short, narrow slit at anterior end. Albumen gland 
lying alongside capsule gland, forming curved mass, 
opening laterally to capsule gland along C-shaped 
opening. Capsule gland weakly glandular, forming 
thin-walled voluminous sac and functioning as brood 
pouch (Fig. 9A?D, bp). Lateral lamina bearing large, 
convoluted, glandular fold along ventral floor, over- 
hanging ventral channel (Fig. 8C, vc); fold tapering 
anteriorly and posteriorly. Medial lamina bearing 
sperm gutter along ventral edge, dorsal spermato- 
phore bursa (spb) and ventral seminal receptacle 
(res). Spermatorphore bursa opening via small ante- 
rior aperture to sperm gutter. Seminal receptacle, 
separated along ventral edge from ventral channel by 
longitudinal flap within inner wall of sperm gutter, 
posteriorly forming blind pouch. Male reproductive 
anatomy unknown. 
Brood pouch in one specimen containing 53 shelled 
juveniles of one whorl or more, and 142 yellowish 
embryos (emb) of less than one whorl. Embryos 
maturing anteriorly within pouch, from fertilized egg 
at posterior end (Fig. lOA) to shelled juveniles of 
approximately 1.5 whorls at anterior end (L). Early 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     379 
Figure 9. Soft body of mature female Potadomoides pelseneeri with uterine brood pouch (paratype IRSNB, no. 145). 
A, B, C, external view. D, detail of brood pouch. Scale bar = 1 mm. bp, brood pouch; dgl, digestive gland; f, foot; go, gonad; 
int, intestine; kDa, kidney; sn, snout; st, stomach. 
embryos displaying typical spiral cleavage (A?D). 
Coiling of visceral mass becoming apparent during 
early organogenesis, with snout and foot beginning to 
differentiate first (E). Cephalic tentacles becoming 
apparent as shell on apex of visceral mass beginning 
to calcify (F). Mantle edge and propodium becoming 
evident next (H), followed by further elaboration of 
snout, tentacles, foot, and operculum (I?K). Process of 
shell formation on apical tip proceeding by secretion 
of small, individual plates that increase in size and 
eventually coalesce (G?K); accretionary shell growth 
commencing within brood pouch at approximately 1 
whorl (L). 
Distribution and ecology 
Known only from the delta of the Malagarasi River at 
Lake Tanganyika (Fig. 1). The only available informa- 
tion on the habitat was given by Leloup (1953: 102) 
who reported living animals to be collected in 
30?40 cm within a calm bay along the river, obtained 
by sieving muddy substrate and vegetable debris with 
a small net. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
380     M. GLAUBRECHT and E. E. STRONG 
Figure 10. Embryos of Potadomoides pelseneeri. A, fertilized egg. B, two-cell stage. C, five-cell stage with four mac- 
romeres and one micromere. D, thirteen-cell stage showing typical spiral cleavage. E, spire down; head-foot beginning to 
differentiate. F, head and snout beginning to form; initial stages of calcification on shell apex by small plate formation. 
G, snout and foot more defined. H, hentacles and mantle edge defined. I, calcification progressing through fusion of plates; 
operculum now present. J, initial shell almost fully fused. K, accretionary growth beginning; snout, foot, cephalic tentacles 
and mantle well developed. L, initial protoconch whorls fully calcified and accretionary growth completed through 
approximately 0.5 whorls. Scale bar = 40 |xm. ap, shell apex; f, foot; me, mantle edge; op, operculum; sn, snout; t, cephalic 
tentacle. 
Remarks 
There are some minor discrepancies between the shell 
measurements found here and in Leloup (1953: 102? 
103), who listed shell parameters (height and width of 
shell and aperture) for 20 specimens for which the 
following averages can be calculated: height = 6.8 ? 
2.94 mm, width = 3.8 ? 1.32 mm. In his text, he is only 
referring to the largest specimen with a shell of 
12 X 5.8 mm. The aperture height and width was 
reported by Leloup with height = 3.6 ? 1.51 mm and 
width = 2.2 ? 0.88 mm, respectively, with 5.9 x 3.6 mm 
for the aperture of the largest shell. This corresponds 
reasonably well with the above given measurements 
for Leloup's complete lot in IRNSB. Brown (1994: 129) 
only mentioned the largest specimen with a shell of 
12 X 6 mm. 
Minor differences in the number of radular denticles, 
in particular of the marginal teeth, are evident when 
comparing with Leloup (1953: fig. 57P), who reported 
up to three outer denticles for the laterals, a 1/1/2 
pattern for the inner marginals and ten denticles for 
the outer ones. Given the scarcity of material avail- 
able, it remains unclear whether these differences 
reflect individual (i.e. intraspecific) variation or obser- 
vational difierences. Similarly, it is not clear whether 
the lower number of denticles {N = 8) found in the 
outer marginals of the adult P. pelseneeri can be con- 
sidered a distinct feature for the type species. However, 
it should be noted that the number of outer cusps 
flanking the broad, spatulate denticle of the laterals in 
P. pelseneeri were also consistently lower (one with an 
occasional second denticle) in the adult than in the 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     381 
juvenile studied herein, a phenomenon also found in 
Lavigeria (M. Glaubrecht, unpubl. data). 
Leloup (1953: 103, fig. 47B) was also the first to 
point out the viviparous reproduction of this species, 
specifying that the anterior part of the oviduct func- 
tions as a brood pouch. Glaubrecht (1996: 148?149, 
pi. 21, figs 6, 7, 8, 9) described and depicted for the 
first time the distinctive wrinkled sculpture in the 
embryonic shell of P. pelseneeri, a feature shared with 
many other brooding limnic cerithioideans, and 
standing in contrast to the shell sculpture of ovipa- 
rous paludomids from Africa. 
POTADOMOIDES BEQUAERTI (DAUTZENBERG & 
GERMAIN, 1914) 
Cleopatra bequaerti Dautzenberg & Germain, 1914: 
59, pi. IV, figs 1, 2, 3, 4, 5, 6 
Potadomoides bequaerti - Mandahl-Barth, 1967: 129, 
pi. 2, figs 11 and 16; Brown, 1994: 130, fig. 62c 
T^pe locality 
DRC: Lualaba River at Kindu (3?S) (Dautzenberg & 
Germain, 1914: 60). 
T^pe material 
Lectotype (Fig. 2B), by present designation, in order 
to enhance the stability of nomenclature in accor- 
dance with article 74.7.3. of the ICZN (MRAC 47.358; 
from 'Congo, Kindu, Lualaba River, stn. 15, 3?S, 
October 30, 1910'; material illustrated in Dautzen- 
berg & Germain, 1914: pi. IV, figs 1, 2, 3, 4, 5, 6). Two 
paralectotypes (MRAC 47.358a, 47.359; from same 
locality). Four paralectotypes (MRAC 47.351-4, local- 
ity as above). Three paralectotj^es (MRAC 47.355- 
357, locality as above) (Fig. 2B). Two paralectotypes 
(MCZ 41020; from 'Belgian Congo, Lualaba River, 
Kindu, 3?S, no. 15; leg. 30 December [sic] 1910, ex 
coll. H. Schouteden; ded. 8 January 1934, ace. 830'). 
The largest shell (8.9 x 6.0 mm) of lot MRAC 
47.358-9 (Fig. 2B), selected here as lectotype, corre- 
sponds well with the original figures (figs 1, 2) and to 
the measurements provided in the original text. 
According to the notes on the labels for lots in MRAC 
and MCZ, both clearly belong to the same original 
syntype series, with those in MCZ erroneously 
labelled as 'paratypes'. Unfortunately, the original 
authors gave no numbers for the originally collected 
sample. Thus, it can only be assumed that a total of 
12 specimens represents the syntype series, of which 
two are in the MCZ and the remaining ten specimens, 
including the lectotype, are in the MRAC. Note that 
one small shell in lot MCZ 41020, with operculum and 
radula, is clearly different and should be separated as 
a distinct species (M. Glaubrecht, unpubl. data). 
Other material examined 
One female found without shell (MRAC 341.980; 
locality Congo, Lualaba at Lokandu, Camp Michita- 
rie) in ethanol, lacking buccal mass; mantle roof sepa- 
rated (Mandahl-Barth, 1967: 129). Oviduct sectioned 
according to methods above. A single shell, found by 
the present authors in Mandahl-Barth's collection at 
the DBL with the same locality data (indicating as 
collector 'Vercammen, Jan 1951'), as well as a radula 
and operculum mounted in Canada Balsam also 
found in Mandahl-Barth's collection at the DBL, are 
most likely from the same animal. 
Taxonomic remarks 
Although Dautzenberg & Germain (1914) originally 
described this species as Cleopatra, we agree with the 
interpretation of the radular features by Mandahl- 
Barth (1967: 129, fig. 16) and follow his assignment of 
the species to Potadomoides. Note that, according to 
the information given by the original authors, the 
type series of P. bequaerti was collected at Kindu in 
October 1910, not December 1910 as noted in error for 
the two paralectotypes in MCZ (41020) (see above). 
Description 
Shell (Fig. 2B); Shell globulose to angular, with four 
basal ridges, varying from those with pronounced 
median (peripheral) keel or cord, contributing to a 
pyramidal shell profile, to those shells with less 
carinated to even rounded whorls. The median carina 
in some specimens forming undulating nodules or 
low spines. Size up to 8.9 x 6.0 mm with 3?4 whorls 
with average parameters for 13 shells: height = 
6.65 ? 1.73 mm, width = 4.80 ? 1.00 mm, height/width 
index = 70.44 ? 9.49 mm (Table 1, Fig. 4). 
Embryonic shell: All specimens were severely eroded 
and embryonic shell not preserved. 
Operculum: Ovate, thin, corneus, translucent; pau- 
cispiral with subcentral spiral nucleus on columellar 
side (MRAC 341.980; MCZ 41020). 
External anatomy: Anterior pedal gland opening to 
short, shallow slit along propodium. Snout large, 
squarish, dorso-ventrally fiattened, transversely 
grooved. Retracted cephalic tentacles approximately 
equal to snout in length. Small eyes located half 
way from base of retracted tentacles on slight 
protuberances. 
Mantle edge smooth. Ctenidium with large trian- 
gular filaments, extending from base of pallial cavity 
to mantle margin, curving around anterior end of 
osphradium. Osphradium approximately two thirds 
length of ctenidium, forming smooth, narrow ridge in 
shallow depression bounded by fiap. Hypobranchial 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
382     M. GLAUBRECHT and E. E. STRONG 
Figure 11. Radula of Potadomoides; light micrographs taken from radulae mounted in Canada Balsam (Danish 
Bilharziasis Laboratory; DEL); compare to scanning electron microscopy photographs in Fig. 6 for other species. Note that 
these are the only existing radula preparations for Potadomoides bequaerti and Potadomoides hirta; thus, their radulae 
were examined and photographed under a compound microscope to verify the congeneric identity as Potadomoides. A, C, 
E, Potadom,oides bequaerti (DEL w/o no.; possibly ex MRAC 341.980). E, D, F, Potadomoides hirta (DEL w/o no.). A, E, 
entire width of radular ribbon. C, D, rachidian and lateral teeth. E, F, inner and outer marginal teeth. Scale bar = 100 |xm. 
gland well developed, forming pendulous transverse 
folds between rectum and ctenidium. 
Radula (Fig. IIA, C, E); Only available radula with 
82 rows. Rachidian teeth broadly triangular, tapering 
to V-shaped base, slightly longer than wide. Cutting 
edge bearing 2/1/2 denticles, with the central denticle 
large, bluntly rounded and flanked on either side by 
an inner accessory denticle that is smaller and rather 
pointed, and an outer upward-curving hook-shaped 
cusp. Lateral teeth bearing 1?2/1/3 smooth denticles. 
Central denticle large,  spatulate, bluntly rounded; 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     383 
Figure 12. Midgut morphology of Potadomoides bequaerti, dorsal view. Midgut opened laterally on the right, roof 
reflected to the left; anterior is uppermost (coarse stippling indicates cuticle lining roof). Scale bar = 1 mm. amf, accessory 
marginal fold; ap, accessory pad; c, caecum; cr, crescentic ridge; cu, cuticle lining stomach roof; e, oesophagus; gp, 
glandular pad; gs, gastric shield; mf, marginal fold; p, crystalline style pocket; sa, sorting area; ss, style sac lip; tl, major 
typhlosole; uf, 'u-shaped' fold. 
flanking denticles also rounded and tapering rapidly 
in size. Inner lateral edge tapering to prominent 
basal, pointed projection. Lateral extensions short, 
approximately equal in length to cusped portion of 
tooth, and wing-shaped. Marginal teeth moderately 
long, with flat shafts, tapering to narrowly rounded 
basal edge. Inner and outer marginal teeth dissimilar. 
Inner marginal teeth with broad shafts bearing large 
shovel-shaped, bluntly rounded main denticle flanked 
by one or two small inner denticles and one small 
outer denticle. Outer marginal teeth with narrow 
shafts bearing ten rounded denticles of similar shape 
and size. 
Foregut: Anterior oesophagus bearing complexly 
folded, glandular buccal pouches; right pouch extend- 
ing slightly farther anteriorly into buccal cavity. 
Pouches contiguous along ventral midline of anterior 
oesophagus, separated anteriorly by smooth, triangu- 
lar field of epithelium in posterior buccal cavity. Each 
pouch bounded by thin, longitudinal fold. 
Midgut: Oesophagus (Fig. 12, e) opening ventrally 
on left side of midgut. Marginal fold (mf) passing 
anteriorly from oesophageal aperture, then curving 
posteriorly bordering right edge of sorting area (sa). 
Sorting area triangular, tapering posteriorly, with 
straight left margin. Accessory marginal fold (amf) 
emerging from left side of oesophageal aperture, 
paralleling marginal fold around posterior tip of 
sorting area; details not visible due to damage of 
specimen. Midgut roof to left of sorting area cuticu- 
larized (cu); epithelium transversely folded anteri- 
orly and forming short, rectangular, flattened pad 
posteriorly. Gastric shield (gs) flattened and weakly 
concave, continuous with cuticle of stomach roof and 
crystalline style pocket (p). Glandular pad (gp) 
large, rounded with small, rounded protuberance 
(ap) at anterior end to left of gastric shield. Cres- 
centic ridge (cr) extending back from oesophagus 
and fusing to right side of glandular pad, bounding 
deep crescentic groove. Only single digestive gland 
duct preserved, opening to deep pocket between 
proximal end of crescentic ridge and glandular pad. 
Shallow caecum (c) extending ventrally under glan- 
dular pad behind gastric shield. Single, fine, longi- 
tudinal fold along midgut floor behind gastric shield, 
opposite caecum. Field of flne parallel folds to right 
of marginal fold extending anteriorly from oesopha- 
gus to base of major typhlosole (tl).  Single broad 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
384     M. GLAUBRECHT and E. E. STRONG 
Figure 13. Reproductive anatomy of Potadomoides bequaerti. Ventral view vsrith semidiagrammatic cross sections 
through gonoduct. Capital letters refer to histological sections in Fig. 15. For simplicity, note that the position of embryos 
in proximal oviduct is not indicated in reconstruction. Scale bar = 1 mm. agl, albumen gland; cgl, capsule gland; 
gf, glandular fold; ovi, renal oviduct; res, seminal receptacle; sg, seminal groove; spb, spermatophore bursa; vc, ventral 
channel. 
fold (uf) along left side of crystalline style pocket 
extending anteriorly to base of pocket below style 
sac lip (ss), bounding U-shaped depression. Style sac 
communicating proximally with intestinal groove. 
Crystalline style present. 
Hindgut: Proximal intestine, lying alongside style 
sac, passing under distal end of style sac behind 
pericardium. Intestine extending posteriorly along- 
side style sac to anterior end of gastric chamber, 
curving under posterior tip of kidney, and continuing 
forward along pallial gonoduct to papillate anus near 
mantle margin. 
Reno-pericardial system: Kidney elongate, oval, 
bounded posteriorly by digestive gland, anteriorly 
overhanging base of mantle cavity and extending into 
pallial cavity alongside gonoduct, ventral to intestine. 
Main chamber containing large mass of excretory 
lamellae separated from ventral bladder alongside 
pericardium. Nephridial gland absent. 
gland weakly glandular (Fig. 14, cgl) and posteriorly 
containing five embryos (emb) of less than one whorl. 
Lateral lamina bearing large, glandular fold (gf) along 
floor, overhanging ventral channel; fold diminishing 
proximally and distally. Medial lamina bearing sperm 
gutter (sg), anterior spermatophore bursa (spb) and 
posterior seminal receptacle (res). Spermatophore 
bursa bearing unorientated sperm (Fig. 140), seminal 
receptacle containing orientated sperm (Fig. 14G). 
Male reproductive anatomy unknown. 
Distribution and ecology 
In addition to the type locality at Kindu, south of the 
confluent of the Elila River into the Lualaba, an 
additional location at Lokandu, north of the Elila, is 
verified based on MRAO material; this location is also 
given by Mandahl-Barth (1967: 129). By contrast 
to the wider ranging P. schoutedeni, this species is 
known only from a relatively short stretch of the 
Lualaba between Lokandu and Kindu (Fig. 1). 
Reproductive system: Renal oviduct forming short, 
recurvent loop before entering base of glandular 
pallial oviduct. Pallial oviduct (Fig. 13), with proximal 
albumen (agl) and distal capsule glands (cgl), opening 
to mantle cavity via short, narrow slit at anterior end. 
Albumen gland lying alongside capsule gland, 
forming 0-shaped tube. Proximal albumen gland com- 
municating with seminal receptacle (res) via narrow 
slit over prominent, dorsally projecting flap. Oapsule 
Remarks 
In their carination and the presence of spines, albeit 
much lower and less pronounced, the shells of 
P. bequaerti resemble P. hirta, as has been already 
noted by Dautzenberg & Germain (1914: 60). Inter- 
estingly, these two taxa are only known from a few 
localities further downstream on the Lualaba River, 
in regions where also the noncarinated P. schoutedeni 
occurs (apparently) sympatrically (Fig. 1). 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     385 
Figure 14. Reproductive anatomy ot Potadomoides bequaerti. A, B, C, D, E, F, G, H, histological cross-sections through 
palhal oviduct from anterior to posterior. Note position of embryos in sections (F, G, H) omitted from reconstruction in 
Fig. 14. Scale bar = 250 |xm. agl, albumen gland; cgl, capsule gland; emb, embryo; gf, glandular fold; hgl, hypobranchial 
gland; int, rectum; kDa, pallial kidney chamber; ovi, renal oviduct; res, seminal receptacle; sg, seminal groove; spb, 
spermatophore bursa; vc, ventral channel. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
386     M. GLAUBRECHT and E. E. STRONG 
Table 1. Shell measurements for four species of Potadomoides and one taxon re-transferred from Potadomoides to 
Cleopatra (M. Glaubrecht, unpubl. data) 
N Height (mm) Width (mm) Height/width x 100 
Potadomoides pelseneeri 55 7.19 ? 1.86 4.72 ? 0.99 66.90 ? 06.30 
Potadomoides bequaerti 13 6.65 ? 1.73 4.80 ? 1.00 70.44 ? 09.49 
Potadomoides hirta 2 10.65 ? 3.46 7.50 ? 0.71 73.22 ? 17.18 
Potadomoides schoutedeni 11 6.17 ? 1.36 4.24 ? 0.76 69.54 ? 08.18 
Cleopatra broecki 3 9.23 ? 1.01 6.90 ? 0.70 74.99 ? 06.01 
Data are the mean ? standard deviation. 
The operculum was unknown to Dautzenberg & 
Germain (1914) but is here verified, based on the 
MCZ and DBL material, to resemble the description 
given for the type species. 
The radula described by Mandahl-Barth (1967) was 
apparently extracted from specimen MRAC 341.980, 
which was found to lack a buccal mass. In contrast to 
the re-description presented herein (Fig. IIA, C, E), 
Mandahl-Barth reported and depicted a lower 
number for several of the small and inconspicuous 
denticles fianking the central rachidian and lateral 
cusp on each side. 
The anatomy for the only animal available was not 
described by Mandahl-Barth (1967). This sample size, 
however, is not suflBcient to establish if the small 
number of young embryos in the pallial oviduct is due 
to immaturity, or if the species has a different repro- 
ductive strategy from that of i-! pelseneeri. In general, 
oviduct morphology corresponds well to that of 
P. pelseneeri. The two differ in the relative positions of 
the spermatophore bursa and seminal receptacle. In 
P. pelseneeri, the dorsal bursa overlies the receptacle 
and the latter extends only a short distance posterior 
to the bursa. In P. bequaerti, the receptacle is prima- 
rily posterior to the bursa and extends almost to the 
base of the pallial oviduct. Other minor differences 
relate primarily to the scale of the oviduct, which is 
distended with embryos in the P. pelseneeri. 
POTADOMOIDES HIRTA (DAUTZENBERG & GERMAIN, 
1914) 
Cleopatra hirta Dautzenberg & Germain, 1914: 59; pi. 
IV, figures 11, 12, 13, 14 
Potadomoides hirta - Mandahl-Barth, 1967: 128; pi. 
2, figure 10; Brown, 1994: 130 
T^pe locality 
DRC: Lualaba River at Nyangwe (4?15'S). 
T^^rpe material examined 
Two syntypes (MRAC 47.432, 47.433; from 'Congo, 
Nyangwe, Lualaba Riv., November 15, 1910, no. 49'; 
illustrated by Dautzenberg & Germain, 1914: pi. IV, 
figs 11, 12, 13, 14). The original authors reported on 
four dry syntypes, of which only two are still in 
existence in the MRAC; the fate of the other two 
remains unclear. The measurements originally pro- 
vided for one of the syntypes (14 x 9 mm, and 
7x6 mm for the aperture) surpassed those found in 
the two existing shells. Thus, we refrain from lecto- 
type designation. 
Other material examined 
The single specimen from Kasango reported by 
Mandahl-Barth (1967: 128), of which the origin and 
source is unknown, was not found in the collection of 
the MRAC (F. Puylaert, pers. comm.), nor is its shell 
present in the DBL. Mandahl-Barth (1967) reported 
on a juvenile shell and radula of hirta which he did 
not illustrate. We assume that the radula mounted in 
Canada Balsam (Fig. IIB, D, E) that we found in 
Mandahl-Barth's collection at the DBL is from the 
same specimen. 
Taxonomic remarks 
Confusion long existed as to the systematic placement 
of hirta. The original authors, Dautzenberg & 
Germain (1914), assigned hir-ta to the genus Cleo- 
patra. In his revision of Potadomoides, Mandahl- 
Barth (1967) assigned this species to the genus, 
apparently based on his study of the radula. However, 
as he provided no illustration, some doubt remained 
as to the suggested assignment. Our re-investigation 
made possible by the specimen found in the DBL 
supports the transfer to Potadomoides. 
Description 
Shell (Fig. 2C): The shells are distinct from other 
congeneric taxa in their knotted ornamentation, 
caused by seven to eight hollow spines projecting from 
elevated ridges or carinae. On the well-rounded ulti- 
mate of three to four whorls, four strong knotted 
spiral cords appear; additional cords are added at the 
base. The periostracum is brown, but partly eroded in 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     387 
the hirta syntypes. Shells have clear growth lines and 
microscopic sculpture of fine spiral striae. The aper- 
ture is ovate, expanded at the lower base near the 
columella which is oflf-white in colour, but not 
calloused. Shell size for the two P. hirta shells 
studied (MRAC 47.432-33) ranges from 8.2 x 7.0 mm 
to 13.1 X 8.0 mm (aperture from 4.6 x 3.3 mm to 
5.0 X 4.0 mm): average measurements given in 
Table 1; see also Figure 4. For the now missing juve- 
nile specimen, Mandahl-Barth (1967: 131) gave a 
shell length of 5.1 mm. 
Radula (Fig. IIB, D, E); Rows 70 (N = 1). Rachidian 
teeth narrowly triangular, tapering to V-shaped base, 
slightly less than 1.5-fold longer than wide. Cutting 
edge bearing 1?2/1/1?2 smooth denticles; outermost 
denticles very small when present. Central denticle 
long, narrowly rounded, with flanking accessory 
denticles smaller and rather pointed. Lateral teeth 
bearing 1/1/3 smooth denticles. Central denticle large, 
spatulate, bluntly rounded; flanking denticles also 
rounded and tapering rapidly in size. Inner lateral 
edge tapering to prominent basal, pointed projection. 
Lateral extensions short, wing-shaped, approximately 
two thirds length of cusped portion of tooth. Marginal 
teeth moderately long, with flat shafts, tapering to 
rounded basal edge. Inner and outer marginal teeth 
dissimilar. Inner marginal teeth with broad shafts 
bearing large shovel-shaped, bluntly rounded main 
denticle flanked by one (to two) small inner denticle(s) 
and one small outer denticle. Outer marginal teeth 
with narrow shafts bearing ten rounded denticles of 
similar shape and size. 
Distribution and ecology 
In addition to the type locality for P. hirta at Nyangwe 
in the Lualaba River (4?15'S), Mandahl-Barth 
(1967: 128) reported this taxon also from Kasongo, 
approximately 70 km further upstream (Fig. 1). 
Remarks 
Reaching a shell size of 14?15 x 9 mm (Mandahl- 
Barth, 1967: 128; Brown, 1994: 130), hirta is the 
largest of the Potadomoides species. With its strong 
ridges and spines P. hirta closely resembles another 
limnic African cerithioidean snail, Potadoma ponthi- 
ervillensis, currently assigned to the Pachychilidae. 
Their shells differ in size but not in proportion and in 
characteristic tuberculate ornamentation. Pilsbry & 
Bequaert (1927: 286) reported and figured juveniles of 
spinose ponthiervillensis with three whorls that reach 
14.7 X 9.3 mm (with apertures of 8.8 mm) which cor- 
responds well with hirta shells. This convergence 
certainly provides a possible source for misidentifica- 
tions of these two taxa in museum collections and the 
field alike, in particular since the known locality of 
P. hirta at the Lualaba River falls within the distri- 
butional range of P. ponthiervillensis. 
However, not only are the distinctly ornamented 
adult shells of ponthiervillensis significantly bigger, 
but also a comparison of the radula clearly allows 
separation of these taxa. Mandahl-Barth (1967: 128) 
reported for a young P. hirta from Kasongo that the 
radula 'is very similar to that of P. schoutedeni', albeit 
without providing either figure or description. The 
mounted radula found in Mandahl-Barth's collection 
at the DBL (see section Material above) in general 
reveals a similar cusp pattern to P. schoutedeni, 
except for the occasional presence of only a single 
outer cusp on the rachidian and the presence of only 
a single inner cusp on the lateral teeth. 
POTADOMOIDES SCHOUTEDENI (DAUTZENBERG & 
GERMAIN, 1914) 
Cleopatra schoutedeni Dautzenberg & Germain, 1914: 
58; pi. IV, figures 15, 16 
Potadomoides schoutedeni - Mandahl-Barth, 1967: 
128, pi. 2, figure 9; Brown, 1994: 130 
Type locality 
DRC:    Nyangwe,    Lualaba    River;    station    50    of 
Dautzenberg    &    Germain    (1914:    58);    as    here 
restricted. 
Type material 
Lectotype, by present designation, in order to enhance 
the stability of nomenclature in accordance with 
article 74.7.3. of the ICZN (MRAC 47.804; 'Nyangwe, 
Lualaba River, November 15, 1910, no. 50'; specimen 
illustrated by Dautzenberg & Germain, 1914: pi. IV, 
figures 15, 16; note that four specimens were 
originally given) (Fig. 3D). One paralectotype (MRAC 
47.805; 'entre Ankoro et Kikondja, Nov 1911, no. 158'; 
originally given by Dautzenberg & Germain, 1914 
with two specimens from being on the Luvua River, 
but is actually at the Lualaba River; Fig. 1) (Fig. 3D). 
One paralectotype (MRAC 47.806: from 'Kindu, 
Lualaba River, 3?S, October 30; no. 16'). One paralec- 
totype (MCZ 41021; 'Lualaba River, Kindu; ex coll. H. 
Schouteden, ded. 8 January 1934, ace. 830'). Five 
paralectotypes (MRAC 47.807-11; 'Luvua River, entre 
Kiambi et Ankoro; leg. Dr Gerard, no. 119'; one speci- 
men used for SEM photograph of embryonic shell in 
Fig. 3G, H, I). Three paralectotypes (MRAC 47.812- 
814; 'Congo, Kibombo, Lualaba River, 4?S, Dec. 1910, 
no. 77'; five juvenile specimens were originally given; 
one specimen with dried soft body was removed from 
shell MRAC 47.812 and radula extracted for SEM 
study) (Fig. 3D). 
Other material examined 
One   specimen   (MRAC   791.851;   'Congo,   Lualaba 
River, Kasongo, leg. R L. G. Benoit, Aug. 8, 1959; det. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
388     M. GLAUBRECHT and E. E. STRONG 
Table 2. Embryonic shell measurements (|xm) of the type species Potadomoides pelseneeri and of Potadomoides schout- 
edeni compared to species of Lavigeria from Lake Tanganyika (M. Glaubrecht, unpubl. data) 
N Width Height Diameter Growth line 
Lavigeria nassa - B 
Lavigeria nassa - K 
Lavigeria littoralis 
Lavigeria livingstoniana 
Vinundu spinulosa 
Lavigeria (mean) 
Potadomoides pelseneeri 
Potadomoides schoutedeni 
12 189.2 ? 17.30 
5 237.5 ? 15.00 
3 156.7 ? 15.28 
4 140.5 ? 04.95 
6 200.0 ? 05.77 
30 192.0 ? 29.53 
6 171.0 ? 14.40 
1 170.0 
114.2 ? 20.98 444 ? 32.3 1.25 
128.0 ? 18.24 597 ? 18.1 1.25 
78.3 ? 14.43 390 ? 34.6 1.20 
84.3 ? 04.60 440 ? 04.1 1.20 
97.8 ? 08.50 498 ? 32.7 1.20 
106.0 ? 22.37 472 ? 67.3 1.20 
73.0 ? 02.40 370 ? 18.0 1.25 
80.0 400 _ 
Data are the mean ? standard deviation. 
B,  population  of Lavigeria  identified  as Lavigeria  nassa  from  Bujumbura  (Burundi);  K,  population from  Kigoma 
(Tanzania). 
Note that species assignments in Lavigeria are preliminary only, pending a comprehensive systematic revision, and that 
V. spinulosa  might be identical with that re-described recently by Michel  (2004)  as  Vinundu guillemei  (Martel & 
Dautzenberg, 1899). 
Mandahl-Barth). Radula mounted in Canada Balsam 
found in Mandahl-Barth's collection at the DBL. 
Description 
Shell (Fig. 2D).- Shell with four or five convex whorls, 
and mostly eight or nine (but occasionally up to 16) 
low, narrow spiral cords per whorl; also visible up to 
eight or nine regular dark bands. Low transverse 
axial ribs present on upper half of whorls, creating 
fine reticulate pattern. Shells ranging in size from 
4.5 X 2.9 mm to 8.9 x 5.6 mm (Fig. 4); average mea- 
surements for 11 specimens are: height = 6.17 ? 
1.36 mm, width = 4.24 ? 0.76 mm, height/width = 
69.54 ? 8.18 mm; Table 1. These parameters corre- 
spond well to the measurement (9x6 mm, aperture 
5.5 X 4 mm) given by the original authors. 
Operculum: Paucispiral with subcentral nucleus. 
tacles approximately equal to snout in length, with 
small eyes located half way from base of retracted 
tentacles on slight protuberances. 
Mantle edge smooth, with light gonoduct, dark 
rectum and ctenidium clearly visible dorsally. No sign 
of genital grove or ovipositor at right head-foot. 
Ctenidium with large triangular filaments, extending 
from base of pallial cavity to mantle margin, curving 
around anterior end of osphradium. Osphradium 
extending approximately two-thirds of the length of 
ctenidium, bounded between two parallel elevated 
ridges. Hypobranchial gland well developed. Rectum 
terminating some distance away from mantle edge, 
near anterior tip of gonoduct. Gonoduct closed with 
slit-like distal opening. Due to the state of preserva- 
tion it could not be resolved whether this specimen is 
male or female (in which case is nonbrooding); poste- 
rior body whorls with gonads were not preserved. 
Embryonic shell (Fig. 3G, H, I); All adult specimens 
eroded, thus details of embryonic shell ornament were 
not preserved with the exception of one juvenile shell. 
Shape pyramidal, with angulated ultimate whorl, 
upper half slightly convex, and pronounced median 
Carina. The size parameters for the only available 
embryonic shell preserved in the juvenile specimen 
(MRAC 47.807-11) corresponds well with the average 
measurements found in P. pelseneeri (Table 2). 
Anatomy: The following description is based on a 
single rehydrated soft body taken from a shell in lot 
MRAC 47.812. Foot with anterior pedal gland opening 
to short, shallow slit along propodium; snout squar- 
ish, dorso-ventrally flattened. Retracted cephalic ten- 
Radula (Fig. 6G, H, I); Two specimens, with up to 
76 rows, studied under SEM and stereomicroscope. 
Rachidian teeth triangular, with broad, convex 
cutting edge tapering to V-shaped base; tooth slightly 
longer than wide; cutting edge bearing 2/1/2 smooth 
denticles (only rarely is there a tiny third, upward- 
curving, outermost denticle); central denticle large, 
bluntly rounded, with flanking accessory denticles 
smaller, somewhat pointed and tapering laterally 
in size. Lateral teeth bearing l(2)/l/2?3 smooth 
denticles; central denticle large, spatulate, bluntly 
rounded, with rounded flanking denticles; inner 
lateral edge straight, tapering along lower half to 
prominent basal, pointed projection; lateral exten- 
sions short, wing-shaped, slightly longer than cutting 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     389 
Table 3. Morphological data matrix used in the phylogenetic analysis for  15 paludomid taxa, including the Asian 
Paludomus, and the thiarid Melanoides as the outgroup 
0 1 2 3 4... 
0112 0111 000110100113 0110221111100100-0?? 
1010 00 000010000013111121101111001101?? 
001?000 10100-04000211110001121100170??? 
0000000 ?00?????000211112101111?00?????? 
1011101200--?01?????0002111100011101001101?? 
1110 0111 00011012 013223211000000010011100 
1110 0112 0111000312 00013323211000000100071100 
10110010 000112 0101032 021100000001001177? 
10111012110070077777010223212000001070011777 
10111012122271077777777220211000001770777777 
10110010 11071201010223212000001770077777 
1000000 10000-010003 02 012 010011000110102 
0011000 00000-00122010012111121100000177 
2000000 00000-0012 0123010112221100110101 
10122011--2-12011103000210012201110010110100 
2 000 000 0 012 0-02 000222 012211121000110110 
Melanoides tuberculata 
Paludomus siamensis 
Cleopatra bulimoides 
Cleopatra ferruginea 
Cleopatra johnstoni 
Lavigeria sp. A 
Vinundu westae 
Potadomoides pelseneeri 
Potadomoides bequaerti 
Potadomoides hirta 
Potadom,oides schoutedeni 
Reymondia horei 
Syrnolopsis lacustris 
Stanleya neritinoides 
Tiphobia horei 
Tanganyicia rufofilosa 
-, inapplicable characters; ?, missing information. 
edge of tooth. Marginal teeth moderately long, with 
flattened shafts, tapering to narrow, rounded basal 
edge. Inner and outer marginal teeth dissimilar. 
Inner marginal teeth with broad shafts bearing large 
shovel-shaped, bluntly rounded main denticle flanked 
by two inner denticles, of which the innermost is 
larger and rounded, and the outer is small and rather 
pointed. Outer marginal teeth with narrow shafts 
bearing 10 rounded denticles of similar shape and 
size, thus having 2/1/1+10. Examination of radula 
mounted in Canada Balsam reveals a pattern con- 
sistent with that observed in scanning electron 
micrographs. 
Distribution and ecology 
Judging from available locality data, P. schoutedeni is 
the most widespread taxon within the genus. In addi- 
tion to type locality at Nyangwe, it was reported from 
Kindu and Kibombo at the Lualaba River, and from 
the Luvua River between Ankoro and Kiambi and 
between Ankoro and Kikondja (Fig. 1). 
Remarks 
In shell shape and size P. schoutedeni most resembles 
the type species P. pelseneeri, although Brown (1994: 
130) compares it to P. bequaerti. The pyramidal shape 
of the juvenile shell (Fig. 3) shows not only some 
resemblance to the pyramidal whorls in P. bequaerti, 
but also to figures of juvenile shells of P. pelseneeri 
provided by Leloup (1953: 103, fig. 47A). According 
to Mandahl-Barth (1967), the radula differs from 
P. bequaerti in bearing two outer cusps on the inner 
marginal teeth and possessing only eight cusps on 
the outer marginals (1/1/2+8). Re-examination of the 
material prepared by Mandahl-Barth revealed the 
tiny outside cusps of the outer marginal teeth by 
inverting the slide and examining the radular ribbon 
from below. The inner marginal teeth were observed 
to only possess a single outer cusp. These observa- 
tions are consistent with the scanning electron micro- 
graphs made on a second specimen (Fig. 6G, H, I). 
PHYLOGENETIC ANALYSIS 
The 44 morphological characters including features 
from the adult and embryonic shell, operculum, 
radula, and reproductive tract used in the phyloge- 
netic analysis are listed in the Appendix. The data 
matrix is given in Table 3. We analysed a total of 15 
paludomid tsLxa including four Potadomoides species, 
seven representative thalassoid taxa, three African 
riverine Cleopatra, and an Asian Paludomus; 
Melanoides is the outgroup. 
Four most parsimonious trees of 128 steps long, 
with consistency index of 0.57, and retention index of 
0.60, were obtained, for which the strict consensus 
tree is shown in Figure 15. This analysis supports the 
basal position of the Asian Paludomus to the African 
riverine and lacustrine taxa and suggests paraphyly 
of Potadomoides species, of which the type species 
P. pelseneeri from the Malagarasi clusters with the 
lacustrine Lavigeria plus Vinundu. Thus, Potado- 
moides is found to be nested within but not basal to 
a   monophyletic   thalassoid   species   flock   of  Lake 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
390     M. GLAUBRECHT and E. E. STRONG 
X 
CD ft n Of 
fe 
CD 
ITl rJ3 
,i4 
?^ 
.s 
m Ti 
ho 4^ 
rn 
H 
01 
,i4 
JTl 
h-l is 
s ^ 
?a i 
m cfl 
T:J t^ ft -4J 
o 
-?3 ^ 
cfl -4-> 
ho m 
Tl !H 
Cfl ^ 
cl Cfl (1) 
<1 Ti 
-M 
(B CD > 
8 
a 
-4J 
-2 K a n CD 
1 
> 
'ED 
o 
Ti 1-1 lO 
CD ^ 
tj rfl 
tfl <D 
!H ?fl 
CJ 
(U rl nj ft 'fi > 
rn 
!H 
't1 
^ 
ft ft 
?a 3 rfl 
cfl ft 
cS J3 cfl -4J 
u-i rJ3 O 
n m 
m 
to s o cfl 
u II H 
0 K a 
T3 
?g 
a crt 
ft d 
d II 
o O ?n 
01 C30 -^ 
TH 
Cfl II CD rfl 
cfl 
Pi 
Cl f 
Pi 
o 
in ^ 
-M 
CD 
Cfl 
pi <D 
10 cfl 
cfl 
iH Pi 
O 
?r-H 
Cfl 
?" Cfl n 
b ft cfl 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     391 
Tanganyika as was proposed earlier. Instead, the tree 
supports monophyly of a fiuviolacustrine clade for 
which the oldest available name is Nassopsinae 
Kesteven, 1903. 
Based on this phylogenetic hypothesis, we find 
several unambiguous synapomorphic characters for 
this fiuviolacustrine clade. They are: paucispiral oper- 
culum [21], rachidian with lateral hooks [26], inner 
marginals with one spatulate cusp [33], oviduct has a 
vaginal opening only [38], and a short spermatophore 
bursa [40]. 
The clade including P. pelseneeri with Lavigeria and 
Vinundu is most significantly unambiguously sup- 
ported by a radula character: the outer marginals 
with up to eight cusps [34]. This grouping is also 
supported by several shell characters, in particular 
sculptural elements [8, 10, 11, 12]. However, these 
provide only ambiguous support, often hampered by 
missing data. By contrast, the sister group relation- 
ship between Lavigeria and Vinundu is supported by 
five characters, namely three embryonic shell charac- 
ters [1, 3, 19], one adult shell character (i.e axial 
elements found on the last whorl [5]), and a scalloped 
mantle margin [22]. 
DISCUSSION 
SYSTEMATIC AFFINITY OF RIVERINE AND LACUSTRINE 
PALUDOMIDAE 
The systematic relationships of and within the thalas- 
soid species fiock of Lake Tanganyika have only 
recently been analysed applying rigorous cladistic 
methods, using molecular sequence date from partial 
cytochrome c oxidase subunit / (CO /) sequences for a 
small subset of the endemic taxa (West & Michel, 
2000; Michel, 2004) and for CO / plus 16S sequence 
data in a more comprehensive analysis (Wilson et al., 
2004). Although the branching patterns of all three 
previous analyses differ rather significantly, all 
support paraphyly of the thalassoid snails; two of 
these (West & Michel, 2000; Wilson etal., 2004) 
support Lavigeria as the most basal offshoot. The 
latter study supported the riverine Cleopatra as sister 
to a clade comprising all remaining Tanganyikan 
taxa, whereas the others placed Cleopatra as sister to 
Reymondia. In addition, all previous analyses sup- 
ported a sister group relationship between the ovipa- 
rous Vinundu and the ovoviviparous Lavigeria clade. 
However, due to lack of material, these studies were 
unable to use molecular data for Potadomoides and 
were consequently silent about the placement of this 
potential fiuviatile sister or its possible ancestry to 
the thalassoid species fiock. 
In comparison to these analyses using molecular 
data, there are some major differences to the phylo- 
genetic reconstruction obtained here from morphol- 
ogy. First of all, the inclusion of all constituent species 
of the riverine Potadomoides has revealed the first 
insight into the phylogenetic relationships among the 
lacustrine and nonlacustrine paludomids of and adja- 
cent to Lake Tanganyika. These relationships had 
been previously discussed based only on circumstan- 
tial evidence (Coulter, 1991; Brown, 1994), and on the 
character evaluation using Hennigian argumentation 
by Glaubrecht (1996). In that analysis, Potadomoides 
and Lavigeria were identified as adelphotaxa sup- 
ported by several synapomorphic features which set 
these two taxa apart from other thalassoid Tanganyi- 
kan and African paludomid snails, namely the pau- 
cispiral operculum (which is unique among thalassoid 
taxa), the uterine brood pouch, a unique radular 
morphology, and several sculptural elements in the 
embryonic shells (Glaubrecht, 1996: 148-150). This 
finding is substantiated here with the new expanded 
analysis. 
Thus, our cladistic analysis of available morpho- 
logical characters supports: (1) the monophyly of a 
clade Nassopsinae including the riverine Potado- 
moides and lacustrine Lavigeria, and (2) the para- 
phyly of Potadomoides reflecting the disjunction of 
its populations from the Malagarasi in the east and 
from the Congo drainage in the west. However, the 
lack of resolution and low bootstrap support within 
this latter clade (Fig. 15) is a consequence of large 
amounts of missing data for P. bequaerti, P. hirta 
and P. schoutedeni. Therefore, we anticipate that it 
will be possible to refine the resolution of this rela- 
tionship if more material becomes available in the 
future. In addition, our analysis supports the exist- 
ence of several lacustrine clades in Lake Tanga- 
nyika that originated from different fiuviatile 
lineages. We discuss below the main implications of 
the phylogenetic analysis, outlining support from 
other lines of evidence or contradictions with exist- 
ing molecular analyses. 
Lacustrine colonizations 
The basal position of the monotypic thalassoid gas- 
tropod Tiphobia horei (Fig. 16) is quite remarkable for 
several reasons. First of all, the fact that it does not 
cluster with other thalassoid gastropods is clearly in 
confiict with the molecular phylogenies that consis- 
tently have found Tiphobia grouped well within this 
species fiock (West & Michel, 2000; Wilson etal., 
2004). Given its occurrence here as basal even in 
relation to the fiuviolacustrine Nassopsinae, repre- 
senting one of the two major invasions of Lake Tan- 
ganyika that is also supported in the molecular 
phylogenies, Tiphobia needs to be further tested with 
additional data as truly representing another inde- 
pendent colonization of the lake. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
392     M. GLAUBRECHT and E. E. STRONG 
'M. tuberculata i 
? P. stamens is 
\ 
T. horei 
^ 
i 
I LT 
? P. schoutedeni 
^^^ p. bequaerti tjlm^ 
? P. hirta 
? 
- P. pelseneeri 
^B? Lavigeria sp. A 
^ ^^^ V. westae 
C. bulimoides 
C. johnstoni 
C. ferruginea 
^^? R. horei 
^f^ T. rufofll: 
S. lacustris 
S. neritinoides 
/ 
4 
\  I 
M 
LT 
^ LT 
^ 
Figure 16. Hypothesis of phylogeny of thalassoid gastropods from Lake Tanganjdka and adjacent rivers. Note the 
repeated occurrence of uterine brooding (square) versus mesopodial brooding (star) in African paludomids, and sub- 
haemocoehc brooding (circle) in thiarids. Riverine taxa are indicated by white bars, namely the Asian Paludomus, four 
Potadomoides species from the Congo River System (C), with Potadomoides pelseneeri endemic to the Malagarasi River 
(M) east of Lake Tanganyika and species of Cleopatra in Africa (A); black bars indicate the thalassoid paludomids endemic 
to Lake Tanganyika (LT). Note the repeated colonization (arrows) of the lake in (1) Tiphobia horei; (2) the Nassopsinae; 
and (3) the remainder of the thalassoid species flock. Note that the type species Cleopatra bulimoides is the most 
widespread in Africa, with Cleopatra ferruginea not yet adequately characterized as an independent southern East African 
species, whereas Cleopatra johnstoni is restricted to Lake Mweru and the Luapula River. 
In addition to the Nassopsinae, the second major 
clade comprises the riverine Cleopatra, found in a 
basal position, and the rest of the thalassoid Tanga- 
nyikan species flock. This clade represents a lacus- 
trine invasion that involves the majority of distinct 
generic taxa from Lake Tanganyika. The type species 
Cleopatra bulimoides that is most widespread in 
Africa clusters together with Cleopatra johnstoni 
which is restricted to Lake Mweru and the Luapula 
River, both adjacent to Lake Tanganyika. The third 
species Cleopatra ferruginea is found in our analyses 
to be more closely related to the rest of the thalassoid 
gastropod species under study here. It should be 
noted, however, that the later taxon is not yet 
adequately characterized as an independent southern 
East African species (Brown, 1994). 
To summarize, molecular and morphological evi- 
dence hints at nonmonophyly of the thalassoid gas- 
tropod species flock as comprised hitherto, but instead 
indicates the repeated and independent colonization 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     393 
of the lake from at least two clades with riverine 
ancestors: the fluviatile genera Potadomoides and 
Cleopatra. We refrain from changing the current clas- 
sification of the taxa involved, awaiting a more 
detailed phylogenetic and possibly comprehensive 
molecular study of more thalassoid taxa and at least 
the type species P. pelseneeri. 
SPECIES COMPOSITION AND ECOLOGY IN 
POTADOMOIDES 
The genus Potadomoides was originally established 
for a single species P. pelseneeri from the Malagarasi 
River (Leloup, 1953). Out of the four species from the 
Congo River system originally described as belonging 
to Cleopatra (Dautzenberg & Germain, 1914) and 
only later assigned to the genus (Mandahl-Barth, 
1967), evidence presented herein confirms the conge- 
neric status of bequaerti, hirta, and schoutedeni. In 
the absence of radular data, Mandahl-Barth (1967) 
only tentatively assigned a fifth species, broecki, to 
Potadomoides. His hesitation was apparently correct 
because newly-available data (M. Glaubrecht, unpubl. 
data) reveals that broecki does not represent a 
member of Potadomoides. Based on conchological and 
radula characters in concert with the outlying geo- 
graphical occurrence at the Aruwimi River in the 
north of the Congo system, it is suggested that broecki 
is a representative of or closely aflBliated with Cleo- 
patra. Accordingly, the distribution of Potadomoides is 
restricted here to the upper (i.e. south-eastern) tribu- 
taries of the Congo drainage, namely the Luvua and 
Lualaba Rivers, as well as east of Lake Tanganyika in 
the Malagarasi River (Fig. 1). 
All known species of Potadomoides inhabit fiuvia- 
tile environments. In the Congo River, these are sur- 
rounded by the lush vegetation typical of a humid, 
tropical forest, whereas at the Malagarasi River delta, 
terrestrial habitats are dominated by a forest/ 
savannah mosaic and dry deciduous savannah wood- 
land (Livingstone, 1975). By contrast, the lacustrine 
sister group Lavigeria + Vinundu in Lake Tanganyika 
lives in a completely different environmental setting. 
Consequently, Brown (1994: 530) saw a fundamental 
difference in ecological adaptation in Lavigeria s.l. 
from the lake proper versus Potadomoides in the 
adjacent rivers of the Congo drainage, implying that 
evolution had been largely driven by the infiuence of 
highly distinct environmental conditions. Unfortu- 
nately, however, it is impossible to more than specu- 
late on possible infiuences of fiuviatile and lacustrine 
habitats, predation, or other selective pressures on, 
for example, different shell morphology (i.e. form and 
ornamentation), operculum shape and radula mor- 
phology. Nevertheless, one striking feature remains 
obvious: although we find a species-rich assemblage 
in ovoviviparous Lavigeria, the oviparous Vinundu 
from lacustrine habitats as well as the riverine Pota- 
domoides, being ovoviviparous at least in part, are 
species-poor. 
Evaluation of morphological characters 
Shell: With the exception of Cleopatra? broecki 
(Glaubrecht, unpubl. data; see above) and possibly 
P. hirta, there are no striking differences in shell 
dimension within the genus (Fig. 4). By contrast, a 
wide spectrum of conchological disparity is found in 
terms of ornamentation. Brown (1994: 129) noted 
that, in this respect, shells of Potadomoides are 
similar to Cleopatra. However, this only holds true if 
the smooth taxa of the former genus are considered 
(i.e. pelseneeri, schoutedeni) because the overall 
degree of conchological variation by far exceeds that 
described for the latter (Brown, 1994: 120-127). 
Although spiral ridges and keels are known from 
shells of some Cleopatra species, spines as in P. hirta 
and the distinctive carinae as in bequaerti are 
unknown to occur in any other Recent African palu- 
domid. Those taxa with smooth or weakly sculptured 
shells with faint spiral elements at the most, namely 
P. pelseneeri and P. schoutedeni, occur further south 
in the fiuviatile regions of the Malagarasi, Luvua and 
upper Lualaba River. By contrast, the two more 
strongly ornamented species with spiral carinae 
(P. bequaerti) and with spines (P. hirta) appear to be 
restricted to locations at the Lualaba River further 
downstream (Fig. 1). The latter species falls out from 
the overall size range of the other congeneric taxa, 
although morphometrics alone does not allow for spe- 
cific differentiation (Fig. 4) in contrast to the distinc- 
tive shell characters described. 
It was this remarkable variety of shell sculpture 
that first led Brown (1994) to state that this range of 
shell forms in Potadomoides 'suggests that their 
ancestors could have possessed the potential for a 
profuse radiation in the young Lake Tanganyika'. 
However, in light of the results from our phylogentic 
analysis, this holds true only for the origin of the 
Lavigeria and Vinundu lineage but not for the other 
thalassoid gastropods. Thus, in conclusion, we can 
state that at least part of the conchological spectrum 
seen today in these lacustrine gastropods might have 
its origin in closely related riverine taxa, but that this 
is dwarfed by the conchological diversity evident in 
the remaining thalassoid species. Potential causes for 
the observed pattern of conspiciously sculptured Pota- 
domoides species occurring further downstream com- 
pared to the smooth species closer to the headwaters 
remains to be evaluated. 
Embryonic shell: Both uterine brooders Potadomoides 
and Lavigeria exhibit a very distinct embryonic shell 
sculpture.  The  wrinkled  ornamentation  caused  by 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
394     M. GLAUBRECHT and E. E. STRONG 
delayed calcification of the apical cap, in concert 
with the narrow and regular reticulate sculpture of 
the subsequent whorls, are unique among African 
limnic Cerithioidea including the thalassoid species 
(Glaubrecht, 1996; M. Glaubrecht, unpubl. data). The 
size of the embryonic shells in Potadomoides (Table 2) 
reach approximately 0.40 mm in diameter at the 
point of sculptural transition between the first shell 
secreted (presumably within the egg capsule) and 
thereafter is surpassed by those in Lavigeria (with a 
diameter between 0.40 and 0.60 mm in some of the 
taxa under study; approximate mean of 0.47 mm). By 
contrast to both, oviparous paludomids, such as Cleo- 
patra, were found to have at this point a diameter of 
1.20 mm (Glaubrecht, 1996: 307-308). Consequently, 
not only the embryonic shell features, in particular 
the distinct ornamentation considered here as syna- 
pomorphic of the two taxa, but also the embryonic 
shell dimensions hint at a sister group relationship of 
Potadomoides and Lavigeria, correlated perhaps with 
the uterine incubation in both. Of interest in this 
context is that embryonic shells of 'Lavigeria' spinu- 
losa, a taxon now assigned preliminarily to Vinundu 
(Michel, 2004), fall well within the range known for 
other ovoviviparous Nassopsinae (Table 2). 
Radula: It is the radula that sets Potadomoides 
clearly apart not only from the African pachychilid 
Potadoma (with its long radula ribbon and quadran- 
gular, ridged rachidian) but also all other thalassoid 
gastropods (Brown & Mandahl-Barth, 1987: 318; 
Glaubrecht, 1996: 148). On the other hand, the simi- 
larities of the radulae oi Potadomoides and Lavigeria, 
that were first noted by Brown (1994: 129) and later 
studied and described in more detail by Glaubrecht 
(1996: 149?150), in conjunction with the distinct 
embryonic shell sculpture as described here (see also 
Glaubrecht, 1996: 149, 308, pi. 21, figs 5, 6, 7, 8, 9), 
support the sister group relationship of these two 
genera. In addition, it is also shared by the oviparous 
Vinundu, giving further indication as to the mono- 
phyly of this clade. 
Alimentary system: The stomach of Potadomoides 
resembles that found for Lavigeria nassa, particularly 
in the shape of the sorting area with its straight left 
margin. Typically in cerithioideans with a tapering 
sorting area, the left margin is highly curved 
(E. E. Strong & M. Glaubrecht, unpubl. data) but, in 
Lavigeria and Potadomoides, the posterior end of the 
sorting area appears truncated, rendering the left 
margin rather straight. Other midgut features that 
are similar in Potadomoides and Lavigeria but are 
shared with other taxa (e.g. Pleuroceridae and/or 
Thiaridae s.s. as well as Paludomidae) are: the broad 
glandular pad,  a weakly concave gastric shield,  a 
crescentic ridge terminating above a shallow caecum, 
a small fold in the style sac pocket that does not 
extend anteriorly to the style sac aperture, and par- 
tially fused typhlosoles. 
Reproductive biology: With the type species P. pelse- 
neeri being a uterine brooder, it was generally as- 
sumed that all congeneric species are (ovo-)viviparous 
(Mandahl-Barth, 1967: 127; Brown, 1994: 129; 
Glaubrecht, 1996: 309). However, only in P. pelseneeri 
did we find evidence for the closed pallial oviduct 
functioning as brood pouch, filled with all ontogenetic 
stages including many shelled juveniles. By contrast, 
in P. bequaerti only, early eggs or embryos were found 
lying in the proximal, glandular part of the oviduct. 
As stated above, this only existing specimen may not 
be a fully mature female. Unfortunately, the only 
available (dessicated) soft body of P. schoutedeni also 
does not allow an unequivocal assessment of repro- 
ductive mode. Thus, positive evidence for uterine 
brooding is to date only available for P. pelseneeri; 
however, definitive evidence for ovoviviparous repro- 
duction might come to light in the future for other 
congeneric species of the Congo River system west of 
Lake Tanganyika. 
This discussion becomes even more important in the 
context of the evolution of brooding in Lavigeria. 
Although is was commonly held that all taxa in the 
latter genus are uterine brooders (Leloup, 1953; Michel 
etal., 1992; Brown, 1994; Glaubrecht, 1994, 1996, 
1999; Michel, 1994), it was only recently revealed that 
some populations actually lack this viviparous mode 
(Michel, 1995; M. Glaubrecht, unpub. data). Accord- 
ingly, it was suggested that this 'tuberculate' clade 
represents a new genus, Vinundu, described by Michel 
(2004) as sister to the viviparous Lavigeria. 
EVOLUTIONAEY IMPLICATIONS OF UTERINE BROODING 
The evolution of direct development and retention of 
eggs and embryos in the female's body (sometimes 
involving direct nutrient transfer) is often found in 
molluscs and other taxa inhabiting freshwater habi- 
tats. This increased parental investment and devel- 
oping appropriate morphological structure for the 
retention of eggs and embryos, generally termed vivi- 
parity, is highly beneficial to protect the progeny 
and has numerous evolutionary implications, as dis- 
cussed, for example, in a phylogenetic framework for 
cerithioidea gastropods by Glaubrecht (1996, 1999, 
2006) and Kohler et al. (2004). 
Uterine brooding in Potadomoides represents a dis- 
tinctive reproductive strategy. Within the thalassoid 
species flock from Lake Tanganyika, Lavigeria only 
shares it with T. horei (Strong & Glaubrecht, 2007). 
This strategy, however, is clearly different from brood- 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     395 
ing in a mesopodial brood pouch as found in Tanga- 
nyicia rufofilosa (Smith, 1880) (Strong & Glaubrecht, 
2002) and different also from incubation in a sub- 
haemocoehc brood pouch, as, for example, in thiarids 
such as Melanoides, Thiara and Tarebia (cf. Glaubre- 
cht, 1996, 1999, 2006; Schiitt & Glaubrecht, 1999) or 
in pachychilids (Kohler etal., 2004); for the contrast- 
ing uterine brooding strategy in lacustrine pachy- 
chilids, see Rintelen & Glaubrecht (2003, 2005). 
Apparently the yolk supply of the egg is the only 
source of embryonic nourishment in Lavigeria species 
and P. pelseneeri because we have no evidence for any 
other nutrient transfer from the female to the devel- 
oping embryos. Consequently, in both taxa, this strat- 
egy should be regarded as lecithotrophic viviparity or 
ovoviviparity. Among freshwater Cerithioidea, it is 
also found, in addition to the taxa under study here, 
in the other known cases of uterine incubation, 
namely in the pachychilid Tylomelania and Pseudo- 
potamis from SE Asia and northern Australia 
(Glaubrecht & Rintelen, 2003; Rintelen & Glaubrecht, 
2003, 2005) and in the paludomid T. horei (for a 
detailed study, see Strong & Glaubrecht, 2007). 
Although another uterine brooder, the Asian pleuo- 
cerid Semisulcospira, remains to be reinvestigated in 
this respect, we found that the fecundity of Lake 
Tanganyika thalassoid gastropods surpasses that of 
other cerithioideans with uterine brood pouches and 
is uniquely correlated with a specific, longitudinal 
arrangement of folds in the brood pouch, which may 
facilitate such high productivity as in Lavigeria and 
Potadomoides. However, comparison of the uterine 
brood pouches in the three African paludomids Tipho- 
bia versus P. pelseneeri and Lavigeria demonstrates 
that they each bear unique features consistent 
with their independent modification for incubation. 
Despite representing functional analogous structures, 
they also share several similarities in organization 
that are interpreted as symplesiomorphies of the 
Lake Tanganyika species flock; for a more detailed 
discussion, see Strong & Glaubrecht (2007). Interest- 
ingly, the delayed calcification recognizable in uterine 
as well as subhaemocoelic brooders among Cerithio- 
idea produced wrinkled embryonic caps not only 
in the paludomids P. pelseneeri, Lavigeria, and T. 
horei, but also in some viviparous south-east Asian 
pachychilids and Thiaridae s.s., albeit this specific 
sculpture feature of the embryonic shell is character- 
istically different in each case. 
Based on the molecular phylogenies available (West 
& Michel, 2000; Michel, 2004; Wilson etal, 2004) as 
well as the results from our preliminary morphologi- 
cal analysis, there is good support for the conclusion 
that uterine brooding has evolved convergently once 
in T. horei and a second time in Potadomoides plus 
Lavigeria. However, in the latter case, our analysis 
does not allow to judge unequivocally on the possibil- 
ity of this particular incubatory strategy being 
homoplastic within Nassopsinae. Given the sister 
group relationship of the uterine brooder P. pelseneeri 
with Lavigeria + Vinundu, and in light of only 
ambiguous evidence for reproductive strategies in 
other Potadomoides species (Fig. 15), we must con- 
sider two equally likely hypotheses. On the one hand, 
uterine brooding could have evolved once in the 
common (and most likely riverine) ancestor of those 
three genera, but was subsequently lost in the lacus- 
trine oviparous Vinundu. On the other hand, brooding 
could have evolved two times independently within 
this clade, once in P. pelseneeri from the Malagarasi 
River, and once in Lavigeria of Lake Tanganyika 
itself, a hypothesis depicted in Figure 16. 
These competing scenarios, whether oviparity in 
Vinundu represents the plesiomorphic state among 
Nassopsinae or a derived state due to secondary loss 
of a viviparous strategy, have important implications. 
Although we now have evidence for the repeated 
evolution of ovoviviparity in several closely related, 
freshwater certhioidean lineages including uterine 
brooders (Glaubrecht, 1999, 2006; Kohler et al., 2004; 
Strong & Glaubrecht, 2007), the reversal from a 
viviparous strategy in P. pelseneeri and Lavigeria to 
the oviparous strategy in Vinundu would be a most 
surprising, not to say conterintuitive, finding. 
In either case, however, it is noteworthy that vivi- 
parity is neither a key innovation in the context of 
the colonization of lacustrine environments, nor an 
intrinsic factor directly correlated with speciation and 
radiation in the Lavigeria flock. This conclusion, as 
anticipated and discussed earlier for freshwater Cer- 
ithioidea in general and Lake Tanganyika paludomids 
in particular (Glaubrecht, 1996, 1999, 2001, 2004, 
2006; Strong & Glaubrecht, 2002, 2007; Kohler et al., 
2004), is in contrast to long-held assumptions as to 
the role of viviparity in the evolution of freshwater 
gastropods (Boss, 1978; Cohen & Johnston, 1987; 
Johnston & Cohen, 1987; Coulter, 1991; Michel et al., 
1992; Michel, 1994). This caveat of attributing evolu- 
tionary importance by overstating the concept of key 
innovations is supported by similar findings in lacus- 
trine as well as riverine species of two other uterine 
brooders that convergently evolved within Cerithio- 
idea, namely Pseudopotamis restricted to two Torres 
Strait Islands (Glaubrecht & Rintelen, 2003) and 
Tylomelania endemic to Sulawesi (Kohler et al., 2004; 
Rintelen & Glaubrecht, 2005). 
EVIDENCE FOR A PROTO-TANGANYIKA AFFINITY OF 
THE MALAGARASI AND CONGO FAUNA 
Approximately 10% of the African freshwater gastro- 
pod fauna has been found to be endemic to one or a 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
396     M. GLAUBRECHT and E. E. STRONG 
few localities compared to an approximately 25% rate 
of endemicity of larger lakes. The Congo Basin as an 
ancient river system adjacent to and connected with 
Lake Tanganyika contains approximately 15% of the 
African freshwater gastropod fauna as endemic 
species (Brown, 1994). Looking into their possible 
ancestral relationships can readily provide important 
clues to be correlated with data on past connections of 
different hydrological systems, as was noted by 
Coulter (1991: 280 flf.) for fishes exhibiting stronger 
afiBnities and sharing greater numbers of species 
between the Congo basin drainage and Lake Tanga- 
nyika than with any other African lake or river. 
The systematic arrangement suggested by our phy- 
logenetic analysis reflects, in a surprising way, the 
highly distinctive distribution pattern of Potado- 
moides species in the Congo River to the west being 
separated by Lake Tanganjdka from the only other 
congeneric species restricted to the Malagarasi River 
in the east (Figs 1, 16). Interestingly, the same com- 
pelling disjunct distribution pattern in Potadomoides 
is also known from some riverine fishes. For example, 
maternally mouthbrooding riverine Schwetzochromis 
and/or Orthochromis, which are rheophilic haplo- 
chromine cichlids (Roberts & KuUander, 1994; De Vos 
& Seegers, 1998; De Vos etal, 2001) occur with five 
species in the Congo River drainage west of Lake 
Tanganyika whereas nine species live in the Malaga- 
rasi drainage (including also the former tributaries 
Rugufu and Luiche); no Orthochromis appear to occur 
in other tributaries of Lake Tangan3dka. Based on 
this iraws-Tanganyikan distribution pattern, De Vos 
& Seeger (1998) hypothesized that ancestors of these 
14 species belonged to the same hydrological system 
that predates the formation of the Tanganjdka Rift. 
Similarly Fryer & lies (1972), Coulter (1991: 236) and 
De Vos et al. (2001) stressed this aflBnity when men- 
tioning that certain apparently relict fish and crusta- 
cean elements present in the Malagarasi River (but 
not elsewhere in the Tanganyika basin) are also found 
in the Congo River, which indicates the former river 
to have been part of the ancient headwaters of the 
Congo that at one time flowed across the proto- 
Tanganyika region. 
By contrast, a hypothesis put forward by Sturm- 
bauer, Verheyen & Meyer (1998) based on preliminary 
mitochondrial sequence data on substrate spawning 
lamprologini flshes holds that lacustrine Lamprologus 
gave rise to riverine taxa that independently colo- 
nized the upper Congo River system and the Mala- 
garasi, thus questioning the monophyletic status of 
these discontinuously distributed riverine taxa. 
However, De Vos etal. (2001) contradict this view in 
light of the recent discovery of a new Lamprologus 
species from the Malagarasi River with stronger 
afiBnity to the six described congeneric species from 
the Congo Basin than to those from Lake Tanganyika. 
Therefore, they again favour a proto-Tanganjdkan 
interconnection of these riverine cichlids from which 
the lacustrine lamprologini species flock subsequently 
originated. 
TOWARDS AN EVOLUTIONARY ECOLOGY OF 
THALASSOID GASTROPODS 
According to traditional views on the origin of the 
Tanganyikan fauna, its current complexity and diver- 
sity evolved as a response to the dramatic environ- 
mental changes during the lakes' formation through a 
variety of new biotopes, e.g. a mosaic of small shallow 
lakes and swamps as the site of adaptation to lacus- 
trine conditions and minor speciation with subse- 
quent isolation from other water bodies through 
physical barriers and habitat preferences. Thus, 
'diversiflcation of the original stock is believed to have 
resulted chiefly from the variety of new biotopes and 
hence opportunities for radiation provided by the lake 
during and after its formation' (Coulter, 1991: 236). 
However, given the rheophilic life habit of the riv- 
erine Potadomoides and the lacustrine habitat of 
Lavigeria and Vinundu species (see above), their 
ecology clearly contradicts assumptions that these 
shallow lake and swamp environments were suitable 
biotopes for the origin and diversification of the Nas- 
sopsinae clade. Instead, these biotopes are preferred 
by pulmonate gastropods, which today play a minor 
role in Lake Tanganyika. Accordingly, an alternative 
hypothesis on the evolutionary ecology of these palu- 
domid snails will be developed in the following. 
Ancient lakes such as Lake Tanganyika are gener- 
ally considered as cradles and centres of evolution 
(Brooks, 1950; Boss, 1978; Coulter, 1991: 236; 
Martens etal., 1994; Rossiter & Kawanabe, 2000) 
inspiring Michel (1994) to ask 'why snails radiate' 
with the implicit or explicit understanding that the 
thalassoid gastropods actually evolved in the lake 
itself. Until recently, it was only rarely questioned 
whether these endemic species flocks indeed repre- 
sent an in situ radiation, rather than the alternative 
hypothesis that lacustrine lineages might represent 
relicts of fiuviatile taxa that only survived in the lake 
(Glaubrecht, 1996, 1999, 2001, 2004, 2006), which 
was subsequently supported by molecular evidence 
(Wilson etal, 2004). 
Focusing largely on lacustrine taxa, fiuviatile taxa 
of adjacent areas were long neglected as key players 
in the evolutionary history of these faunas; this is 
equally true for studies of other lacustrine species, 
and not just molluscs. Interestingly, there are at least 
a few early but long overlooked statements hinting 
into the opposite direction. For example, Dautzenberg 
& Germain (1914: 2), on discussing the striking con- 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     397 
chological features of the thalassoid or halolimnic 
gastropods in African rift systems (i.e. Lakes Tanga- 
nyika and Malawi) and also Lake Mweru, very early 
on pointed out explicitly: (1) that the existence of 
marine-like shells 'enleve encore d'avantage au Tan- 
ganjdka son charactere d'exception', assuming that 
the analogues milieux of the lake led to a convergent 
evolution of these peculiar forms and (2) regarded 
them actually as representing members of an ancient 
African fauna. As these authors anticipated, this lake 
with its peculiar hydrological features, which in many 
respects provides similar environments as found in 
the ocean, forced ancient African faunistic elements 
colonizing the lake to adapt accordingly. Thus, in 
contrast to the century long held belief in an excep- 
tional endemic radiation of the Tanganyika mollusc 
fauna (Yonge, 1938; Brook, 1950; Boss, 1978; Michel, 
1994), Dautzenberg & Germain (1914) already indi- 
rectly suggested a prelacustrine origin of the thalas- 
soid gastropods of Lake Tanganyika, consequent to 
the discovery of the strongly sculptured species now 
placed in Potadomoides from the Lualaba River. Van 
Damme & Pickford (2003) correctly pointed out that it 
was very unfortunate that for unknown reason little 
attention has been given to the possible relation 
between the Tanganjdkan endemics and these Pota- 
domoides species, with the notable exception of 
Brown (1994) and Glaubrecht (1996). 
Nearly a century later, based on molecular research 
that helped to uncover the existence of several ances- 
tral lineages of cichlid fishes in Lake Tanganyika, 
Nishida (1991) suggested that these lineages might be 
even older than the lake. Furthermore, Banister & 
Clarke (1980), and subsequently Martens (1997), 
noted that Lake Tanganyika might have served as an 
evolutionary reservoir. For the thalassoid gastropods 
Wilson et al. (2004) tested this refuge idea and found 
conclusive evidence, based on a molecular clock 
approach, that this ancient lake is indeed more likely 
a reservoir of ancient gastropods lineages than a 
cradle and centre of escalatory change as suggested 
traditionally by Michel et al. (1992), Michel (1994), 
and West & Cohen (1996). 
Consistent with all available evidence, it is possible 
to provide the following scenario as an alternative 
hj^othesis on the evolution of Tanganyikan gastro- 
pods. Most likely, the ancestors of the Nassopsinae 
were originally members of a river-dwelling palaeo- 
fauna of the Congo basin region, perhaps even part of 
a more widespread ancient malacofauna (Van Damme 
& Pickford, 2003). Potadomoides, surviving in these 
riverine habitats until today, should be considered a 
relic of this primordial stock. Provided our hypothesis 
of Lake Tanganyika as an evolutionary reservoir is 
correct (Wilson et al., 2004), this rift lake, after its 
formation in the late Miocene not only offered a stable 
habitat, but also provided new challenges. Some of 
the riverine members might have entered the lake 
graben, and apparently this inital colonization event 
must have occurred at least two times in ancestors of 
the thalassoid gastropods according to the phyloge- 
netic evidence discussed above. The new evolutionary 
opportunities available to the ancestors of the lacus- 
trine Nassopsinae, resulted in a species fiock of 
closely allied and phenotypically variable members, 
namely the oviparous Vinundu species and the ovovi- 
viparous Lavigeria. Both diverged in lacustrine isola- 
tion and adapted to various habitats in the lake in the 
course of an in situ radiation in this case. 
It remains to be studied in more detail whether the 
evolution of ovoviviparity in P. pelseneeri must be 
considered an independent event that occurred in a 
habitat of the former Congo River headwater that 
later became the Malagarasi River delta, as it is 
suggested from the hypothesis favoured in Figure 16. 
In this case, uterine incubation evolved in parallel 
once in Lavigeria species after the formation of Lake 
Tanganyika, and thus under lacustrine conditions, 
and a second time in P. pelseneeri under riverine 
conditions after the isolation of the Malagarasi River. 
To summarize, both the Congo River drainage 
and Lake Tanganyika provide stable evolutionary 
systems, both in their own right and with distinct 
inherent environmental dimensions. With emphasis 
starting to shift recently to nonlacustrine settings in 
studies of gastropods (Glaubrecht & Kohler, 2004) 
and cichlid fishes (Joyce etal., 2005; Kocher, 2005), 
intralacustrine radiations are not to be seen as evo- 
lutionary dead ends. By contrast, it can be expected 
that the potentially repeated switch of limnic taxa 
between riverine and lacustrine habitats during the 
course of their evolution guaranteed their survival 
and, at the same time, provided ample evolutionary 
opportunities for speciation and radiation under 
various conditions. 
CONCLUSIONS 
Based on cladistic analyses of morphological data, 
we found a sister group relationship of lacustrine 
species of Lavigeria + Vinundu endemic to Lake 
Tanganyika and the fluviatile Potadomoides from 
the Congo drainage system, with P. pelseneeri from 
the Malagarasi River being the closest relative. 
Thus, by contrast to earlier assumptions, Potado- 
moides is not ancestral to the endemic species flock 
of thalassoid Tanganjdkan gastropods. Instead, this 
fluviolacustrine clade, uniting two lacustrine taxa 
and the fluviatile Potadomoides, apparently repre- 
sents an early independent lineage of East African 
paludomids that, until recent times, survived in 
and   adjacent   to   Lake   Tanganyika.   Consequently, 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
398     M. GLAUBRECHT and E. E. STRONG 
although Potadomoides certainly does not serve as a 
suitable model for the majority of the thalassoid 
gastropods, it might serve as a model for an African 
paludomid with the potential to colonize not only 
riverine, but also lacustrine habitats. Despite being 
demonstrably of considerable interest for evolution- 
ary biology, no large-scale exploration of limnic habi- 
tats in the Congo Basin has been undertaken for 
several decades. However, equal to great lakes such 
as Tanganyika and Malawi, the Congo River and its 
easternmost tributaries have an outstandingly rich 
species assemblage, not only of unique molluscs, 
that deserve more attention in the future. 
ACKNOWLEDGEMENTS 
We are indebted to F. Puylaert (MRAC, Tervuren), Dr 
J. Van Goethem and Dr Claude Massin (IRSNB), 
Adam Baldinger (MCZ) and Dr Thomas Kristensen 
(DBL) for loan of the material listed above and for 
their permission to study the anatomy by dissection 
and histology, and for additional information concern- 
ing the t3^e material. We thank Thomas Kristensen 
and Aslak Joergensen for their assistence during our 
visit to the DBL in April 2002, helping us to trace 
down the missing radula slides from Mandahl-Barth 
that allowed verification of generic allocation. We 
gratefully acknowledge the helpful comments of two 
reviewers, to further improve the manuscript. An 
Ernst Mayr Research Grant to M.G. enabled a visit to 
the collection of the MCZ in December 1998. This 
work was supported by grant GL 297/5-1 of the 
Deutsche Forschungsgemeinschaft. 
REFERENCES 
Banister KE, Clarke MA. 1980. A revision of the large 
Barbus (Pisces, Cyprinidae) of Lake Malawi with a recon- 
struction of the history of the southern African Rift Valley 
lakes. Journal of Natural History 14: 483?542. 
Boss KJ. 1978. On the evolution of gastropods in ancient 
lakes. In: Fretter V, Peake J, eds. Pulmonates, Vol. 1. 
London: Academic Press, 385?428. 
Bouchet P, Rocroi JP. 2005. Classification and nomenclator 
of gastropod families. Malacologia 47: 1?397. 
Brooks JL. 1950. Speciation in ancient lakes. Quarterly 
Review of Biology 25: 131-176. 
Brown DS. 1994. Freshwater snails of Africa and their 
medical importance, 2nd edn. London: Taylor & Francis. 
Brown DS, Mandahl-Barth G. 1987. Living molluscs of 
Lake Tanganyika: a revised and annotated list. Journal 
of Conchology 32: 305-327. 
Cohen AS, Johnston M. 1987. Speciation in brooding and 
poorly dispersing lacustrine organisms. Palaios 2: 426?435. 
Coulter GW. ed. 1991. Lake Tanganyika and its life. Oxford: 
Oxford University Press. 
Danley PD, Kocher TD. 2001. Speciation in rapidly diverg- 
ing systems: lessons from Lake Malawi. Molecular Ecology 
10: 1075-1086. 
Dautzenberg P, Germain L. 1914. Recoltes malacologiques 
du Dr J. Bequaert dans le Congo Beige. Revue Zoologique 
Africaine 4: 1-74. 
De Vos L, Seegers L. 1998. Seven new Orthochromis species 
(Teleostei: Cichlidae) from the Malagarasi, Luiche and 
Rugufu basins (Lake Tanganyika drainage), with notes on 
their reproductive biology. Ichthyological Explorations of 
Freshwaters 9: 371-420. 
De Vos L, Seegers L, Taverne L, Thys van den Aude- 
naerde D. 2001. L'ichthyfaune du bassin de la Malagarasi 
(Systeme du Lac Tanganyika): une synthese de la connais- 
sance actuelle. Annales Museum Royal L'afrique Centrale 
(Zoologie) 285: 117-152. 
Fryer G. 1991. Comparative aspects of adaptive radiation 
and speciation in Lake Baikal and the great African rift 
lakes of Africa. Hydrobiologia 211: 137-146. 
Fryer G. 1996. Endemism, speciation and adaptive radiation 
in great lakes. Environmental Biology of Fishes 45: 109? 
131. 
Fryer G, lies TD. 1972. The cichlid fishes of the Great Lakes 
of Africa: their biology and evolution. Edinburgh: Oliver & 
Boyd. 
Gautier A. 1970. Fossil fresh water Mollusca of the Lake 
Albert-Lake Edward Rift (Uganda). Annales Musee Royal de 
LAfrique Centrale, Tervuren, Serie 8, Sciences Geologiques 
67: 1-160. 
Glaubrecht M. 1994. Explosive Evolution in ostafrikanis- 
chen Seen. Entstehung von Artenschwarmen bei Fischen 
und Schnecken. Biologie in unserer Zeit 24: 45?54. 
Glaubrecht M. 1996. Evolutionsokologie und Systematik am 
Beispiel von SUfi- und Brackwasserschnecken (Mollusca: 
Caenogastropoda: Cerithioidea): Ontogenese-Strategien, 
Paldontologische Befunde und Historische Zoogeographie. 
Leiden: Backhuys Publishers. 
Glaubrecht M. 1999. Systematics and evolution of viviparity 
in tropical freshwater gastropods (Cerithioidea: Thiaridae 
S.I.). Courier Forschungs-Institut Senckenberg 203: 91?96. 
Glaubrecht M. 2001. An evolutionary ecology approach to 
the presumptive and endemic radiation of freshwater 
Cerithioidea (Mollusca, Gastropoda) in East Africa and on 
Sulawesi. In: Abstracts world congress of malacology, 
Vienna August 2001: 122. Vienna: Unitas Malacologica. 
Glaubrecht M. 2004. A new look at old lineages: riverine 
ancestry and the origin of adaptive radiations in limnic 
gastropods. In: Abstract of the World Congress of Malacol- 
ogy, Perth, Western Australia, July 2004. Perth: Unitas 
Malacologica, 51. 
Glaubrecht M. 2006. Independent evolution of reproductive 
modes in viviparous freshwater Cerithioidea (Gastropoda, 
Sorbeoconcha) ? a brief review. Basteria 3 (SuppL): 23?28. 
Glaubrecht M, Kohler F. 2004. Radiating in a river: sys- 
tematics, molecular genetics and morphological differentia- 
tion of viviparous freshwater gastropods endemic to the 
Kaek River, central Thailand (Cerithioidea, Pachychilidae). 
Biological Journal of the Linnean Society 82: 275?311. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     399 
Glaubrecht M, Rintelen T. 2003. Systematics, molecular 
genetics and historical zoogeography of the viviparous 
freshwater gastropod Pseudopotamis (Cerithioidea, 
Pachychilidae): a relic on the Torres Strait Islands, 
Australia. Zoologica Scripta 32: 415^35. 
Glaubrecht M, Strong EE. 2004. Spermatophores of thalas- 
soid gastropods (Paludomidae) in Lake Tanganyika, East 
Africa, with a survey of their occurrence in Cerithioidea: 
functional and phylogenetic implications. Invertebrate 
Biology 123: 218-236. 
Johnston MR, Cohen AS. 1987. Morphological divergence in 
endemic gastropods from Lake Tanganyika: implications for 
models of species flock formation. Palaios 2: 413?425. 
Joyce DA, Lunt DH, Bills R, Turner OF, Katongo C, 
Duftner N, Sturmbauer C, Seehausen O. 2005. An 
extant cichlid fish radiation emerged in an extinct 
Pleistocene lake. Nature 435: 90-95. 
Kocher TD. 2004. Adaptive evolution and explosive specia- 
tion: the cichlid model. Nature Reviews Genetics 5: 288?298. 
Kocher TD. 2005. Ghost of speciation past. Nature 435: 
29-30. 
Kohler F, Rintelen T, Meyer A, Glaubrecht M. 2004. 
Multiple origin of viviparity in Southeast Asian gastropods 
(Cerithioidea: Pachychilidae) and its evolutionary implica- 
tions. Evolution 58: 2215-2226. 
Leloup E. 1953. Exploration hydrobiologique du Lac Tanga- 
nika (1946-1947). Resultats scientifiques. Gasteropodes. 
Bruxelles: Institut Royal des Sciences Naturelles de 
Belgique. 
Livingstone DA. 1975. Late Quaternary climatic change in 
Africa. Annual Revue in Ecology and Systematics 6: 249? 
280. 
Lydeard C, Holznagel WE, Glaubrecht M, Ponder WF. 
2002. Molecular phylogeny of a circum-global, diverse 
gastropod superfamily (Cerithioidea: MoUusca: Caeno- 
gastropoda): pushing the deepest phylogenetic limits of 
mitochondrial LSU rDNA sequences. Molecular Phylogenet- 
ics and Evolution 22: 399^06. 
Mandahl-Barth G. 1967. Revision of the African genera 
Potadoma Gray and Potadomoides Leloup, and description 
of a new species of Cleopatra (Gastropoda Prosobranchia: 
Melaniidae). Revue de Zoologie et de Botanique Africaines 
76: 110-131. 
Martens K. 1997. Speciation in ancient lakes. Ti'ends in 
Ecology and Evolution 12: 177-182. 
Martens K, Coulter G, Goddeeris B. 1994. Speciation 
in ancient lakes ? 40 years after Brooks. In: Martens K, 
Goddeeris B, Coulter G, eds. Speciation in ancient lakes. 
Stuttgart: Schweizerbart'sche-Verlagsbuchhandlung, 75?96. 
Michel E. 1994. Why snails radiate: a review of gastropod 
evolution in long-lived lakes, both Recent and fossil. In: 
Martens K, Goddeeris B, Coulter G, eds. Speciation in 
Ancient Lakes. Advances in Limnology, Vol. 44. Stuttgart: 
Schweizerbart'sche Verlagsbuchhandlung, 285?317. 
Michel E. 1995. Evolutionary diversification of rifl; lake gas- 
tropods: morphology, anatomy, genetics and biogeography of 
Lavigeria (Mollusca: Thiaridae) in Lake Tanganyika. PhD 
Thesis, University of Arizona. 
Michel E. 2000. Phylogeny of a gastropod species fiock: 
exploring speciation in Lake Tanganyika in a molecular 
framework. In: Rossiter A, Kawanabe H, eds. Biology of 
Ancient Lakes. Advances in Ecological Research, Vol. 31. 
London: Academic Press, 275?302. 
Michel E. 2004. Vinundu, a new genus of gastropod 
(Cerithioidea: 'Thiaridae') with two species from Lake 
Tanganyika, East Africa, and its molecular phylogenetic 
relationships. Journal of Molluscan Studies 70: 1? 
19. 
Michel E, Cohen AS, West K, Johnston MR, Kat PW. 
1992. Large African lakes as natural laboratories for evo- 
lution: examples from the endemic gastropod fauna of Lake 
Tanganyika. Mitteilungen der Internationalen Vereinigung 
fur Limnologie 23: 85-99. 
Nation JA. 1983. A new method using hexamethyldisilazane 
for the preparation of soft insect tissue for scanning electron 
microscopy. Stain Technology 55: 347?352. 
Nishida M. 1991. Lake Tanganyika as an evolutionary 
reservoir of old lineages of East African cichlid fishes: 
inferences from allozyme data. Experientia 47: 974?979. 
Nixon KC. 1999-2002. Winclada, Version 1.00.08. Published 
by the author. L. H. Bailey Hortorium. Ithaca, NY: Cornell 
University. 
Pelseneer P. 1886. Notice sur les moUusques recueillis par 
M. le Capitaine Storms dans le region du Tanganyka. 
Bulletin Du Musee Royal d'Histoire Naturelle de Belgique 
4: 102-128. 
Pilsbry HA, Bequaert J. 1927. The aquatic moUusks of the 
Begian Congo. With a geographical and ecological account 
of Congo malacology. Bulletin of the American Museum of 
Natural History 53: 69-602. 
Rintelen T, Glaubrecht M. 2003. New discoveries in old 
lakes: three new species of Tylomelania Sarasin & Sarasin, 
1897 (Gastropoda: Cerithioidea: Pachychilidae) from the 
Malili lake system on Sulawesi, Indonesia. Journal of 
Molluscan Studies 69: 3?17. 
Rintelen T, Glaubrecht M. 2005. The anatomy of an 
adaptive radiation: a unique reproductive strategy in the 
endemic freshwater gastropod Tylomelania (Cerithioidea: 
Pachychilidae) on Sulawesi, Indonesia, and its bio- 
geographic implications. Biological Journal of the Linnean 
Society 86: 513-542. 
Rintelen T, Wilson AB, Meyer A, Glaubrecht M. 2004. 
Escalation and trophic specialization drive adaptive radia- 
tion of freshwater gastropods in ancient lakes on Sulawesi, 
Indonesia. Proceedings of the Royal Society of London Series 
B, Biological Sciences 271: 2841-2849. 
Roberts TR, KuUander SO. 1994. Endemic cichhd fishes of 
the Fwa River, Zaire: systematics and ecology. Ichthyologi- 
cal Explorations of Freshwaters 5: 97?154. 
Rossiter A, Kawanabe H, eds. 2000. Biology of ancient 
lakes. Advances in ecological research. Vol. 31. London: 
Academic Press. 
Salzburger W, Meyer A. 2004. The species flocks of East 
African cichlid fishes: recent advances in molecular phylo- 
genetics and population genetics. Naturwissenschaften 91: 
277-290. 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401 
400     M. GLAUBRECHT and E. E. STRONG 
Salzburger W, Meyer A, Baric S, Verheyen E, 
Sturmbauer C. 2002. Phylogeny of the Lake Tanganyika 
cichlid species flock and its relationship to the Central and 
East African Haplochromine cichlid fish faunas. Systematic 
Biology 51: 113-135. 
Schiitt S, Glaubrecht M. 1999. Thiara amarula (Linne, 
1758) (Caenogastropoda: Thiaridae) in Australia ? new 
evidence on the anatomy of the reproductive system in a 
viviparous freshwater mollusc. Courier Forschungs-Institut 
Senckenberg 215: 181-188. 
Streelman JT, Danley PD. 2003. The stages of vertebrate 
evolutionary radiation. Trends in Ecology and Evolution 18: 
126-131. 
Strong EE, Glaubrecht M. 2002. Evidence for convergent 
evolution of brooding in a unique gastropod from Lake 
Tanganyika: anatomy and afEnity of Tanganyicia rufofilosa 
(Smith, 1880) (Caenogastropoda, Cerithioidea, Paludomi- 
dae). Zoologica Scripta 31: 167?184. 
Strong EE, Glaubrecht M. 2003. Anatomy and systematic 
afBnity of Stanleya neritinoides (Smith, 1880), an enigmatic 
member of the thalassoid gastropod species flock from Lake 
Tanganyika, East Africa (Cerithioidea: Paludomidae). Ada 
Zoologica 84: 249-265. 
Strong EE, Glaubrecht M. 2007. The morphology and inde- 
pendent origin of ovoviviparity in Tiphobia and Lavigeria 
(Caenogastropoda: Cerithioidea: Paludomidae) from Lake 
Tanganyika. Organism, Diversity and Evolution 7: 81? 
105. 
Sturmbauer C. 1998. Explosive speciation in cichlid flshes 
of the African Great Lakes: a dynamic model of adap- 
tive radiation. Journal of Fish Biology 53 (Suppl. A): 18? 
36. 
Sturmbauer C, Verheyen E, Meyer A. 1998. Mitochondrial 
phylogeny of the Lamprologini, the major substrate spawn- 
ing lineage of cichlid fishes from Lake Tanganyika in 
Eastern Africa. Molecular Biology and Evolution 11: 691? 
703. 
Van Damme D. 1984. The freshwater mollusca of Northern 
Africa. Distribution, biogeography and palaeoecology. 
In: Developments in hydrobiology. Vol. 25. Dordrecht: Junk. 
Van Damme D, Pickford M. 2003. The late Cenozoic 
Thiaridae (Mollusca, Gastropoda, Cerithioidea) of the 
Albertine Rift Valley (Uganda-Congo) and their bearing 
on the origin and evolution of the Tanganyikan thalassoid 
malacofauna. Hydrobiologia 498: 1?83. 
Verheyen E, Salzburger W, Snoeks J, Meyer A. 2003. 
Origin of the superflock of cichlid fishes from Lake Victoria, 
East Africa. Science 300: 325-329. 
West K, Cohen A. 1996. Shell microstructure of gastropods 
from Lake Tanganyika, Africa: adaptation, convergent 
evolution, and escalation. Evolution 50: 672?681. 
West K, Michel E. 2000. The dynamics of endemic diversi- 
fication: molecular phylogeny suggests an explosive origin of 
the thiarid gastropods of Lake Tanganyika. In: Rossiter A, 
Kawanabe H, eds. Ancient lakes: Biodiversity, ecology and 
evolution. San Diego, CA: Academic Press, 331?354. 
Wilson AB, Glaubrecht M, Meyer A. 2004. Ancient lakes as 
evolutionary reservoirs: evidence from the thalassoid gastro- 
pods of Lake Tanganyika. Proceedings of the Royal Society of 
London Series B, Biological Sciences 271: 529?536. 
Yonge  CM.   1938.  The  prosobranchs  of Lake  Tanganyika. 
Nature 142: 464-466. 
APPENDIX 
LIST OF CHAEACTERS AND CHARACTER STATES 
0. Shell   shape   defined  by   apical   angle:   0 - long 
spire, 1 - medium spire, 2 - short spire; 
1. Early whorls  with   axial  elements:   0 - absent, 
1 - present; 
2. Early whorls with  spiral elements:  0 - absent, 
1 - present; 
3. Form of juvenile whorls: 0 - convex, 1 - angulate, 
1 - stepped; 
4. Form of last adult whorl: 0 - convex; 1 - angu- 
late, 2 - stepped; 
5. Last   whorl   with   axial   elements:   0 - absent, 
1 - present; 
6. Last   whorl   with   spiral   elements:   0 - absent, 
1 - present; 
7. Last whorl spiral elements:  0 - faint  striation, 
1 - low ridges, 2 - elevated carina; 
8. Number  of carinae:  0 - one  carina,   1 - several 
carinae; 
9. Ornament on median carina: 0 - fused, 1 - nodu- 
lose, 2 - spinose; 
10. Ornament on subsutural carina: 0 - fused, 
1 - nodulose, 2 - spinose; 
11. Ornament on other carinae: 0 - fused, 1 - nodu- 
lose, 2 - spinose; 
12. Juvenile peristome: 0 - holostome, 1 - slightly 
siphonate; 
13. Adult peristome: 0 - simple, 1 - slightly sipho- 
nate, 2 - siphonate; 
14. Adult umbilicus: 0 - absent, 1 - present; 
15. Initial cap of shell: 0 - smooth, 1 - completely 
wrinkled, 2 - laterally wrinkled; 3 - granulose; 
16. Embryonic shell with spiral sculpture: 0 - absent, 
1 - present; 
17. Embryonic shell spiral sculpture: 0 - several low 
keels, 1 - one pronounced keel, 2 - faint striation; 
18. Embryonic shell with axial sculpture: 0 - absent, 
1 - present; 
19. Embryonic shell size at first whorl: 0 - < 300 |im, 
1 - 301-400 |im, 2 - 401-600 |im, 3 - 600- 
1000 i^m, 4 - > 1000 )im; 
20. Operculum shape: 0 - ovate, 1 - circular; 
21. Operculum coiling: 0 - concentric with paucispi- 
ral nucleus, 1 - paucispiral, 2 - multispiral; 
22. Mantle edge: 0 - smooth, 1 - papillate, 2 - 
fringed, 3 - scalloped; 
23. Radular ribbon rows: 0?50 or less, 1 - 50?60 
rows, 2 - 60?90 rows, 3 - > 90 rows; 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367^01 
EVOLUTION OF POTADOMOIDES     401 
24. Rachidian shape: 0 - wider than high, 1 - qua- 
drangular, 2 - higher than wide; 
25. Rachidian lateral sides: 0 - straight and vertical, 
1 - convex,   2 - straight  and  at  positive   angle, 
3 - straight and at negative angle; 
26. Rachidian lateral margin: 0 - absent lateral den- 
ticle, 1 - lateral hook-like cusp; 
27. Rachidian basal projection: 0 - rounded, 1 - 
triangular; 
28. Central dentition on cutting edge: 0 - numerous 
(? 10) of slightly descending size, 1 - one main 
middle denticle flanked by one smaller denticle, 
2 - one main middle  denticle flanked by more 
than one smaller denticles; 
29. Laterals: 0 - with short lateral extensions (less/ 
equal to cutting edge), 1 - with long lateral exten- 
sions (up to twice the cutting edge), 2 - very long 
extensions (more than twice the cutting edge); 
30. Basal tongue of lateral: 0 - thickened area 
running into basal projection, 1 - not thickened; 
31. Denticles of lateral cutting edge: 0 - one major 
spatulate, 1 - one major pointed ? same size; 
32. Dentition of lateral: 0 - with one main cusp and 
few (1?3) outer flanking denticles, 1 - with one 
main cusp and >3 outer flanking denticles, 2 - 
with numerous denticles of more or less equal 
size; 
Inner marginals:  0 - one  spatulate  main  cusp, 
1 - with   many   cusps   (up   to   eight),   2 - with 
numerous cusps (>8); 
Outer  marginals:   0 - with  many  cusps  (up  to 
eight), 1 - with numerous cusps (>8); 
35. Ovipositor: 0 - absent, 1 - present; 
36. Uterine brood pouch: 0 - absent, 1 - present; 
Subhaemocoelic      brood      pouch:       0 - absent, 
1 - present; 
Gonoduct aperture: 0 - restricted to tip, 1 - one- 
third length of gonoduct; 
Albumen gland: 0 - straight, 1 - coiled; 
Spermatophore   bursa:   0 - extends   to   base   of 
1 - extends to proximal end of 
33 
34 
37. 
38. 
39. 
40. 
albumen gland, 
albumen gland; 
41. Spermatophore 
1 - present; 
42. Spermatophore: 0 - bifld, 1 - sac-like; 
43. Spermatophore spines: 0 - absent, 1 - 
terior edge, 2 - along lateral edge. 
forming     organ:      0 - absent. 
along pos- 
? 2007 The Linnean Society of London, Biological Journal of the Linnean Society, 2007, 92, 367?401