The Veliger49(3):l?2-139 (October 1, 2007) 
THE VELIGER 
? CMS, Inc., 2006 
Fallacies Underlying the Assumption of Calcium Limitation on the 
Evolution of Land Snails in Bermuda 
STORRS L. OLSON 
Division of Birds, National Museum of Natural History, P.O. Box 37012, Smithsonian Institution, 
Washington, D. C. 20O?3-70I2 
(e-mail: olsons@si.edu) 
PAUL J. HEARTY 
School of Earth and Environmental Sciences, University of Wollongong, Wollongong, NSW 2522, Australia 
Abstract. The supposed lack of calcium during glacial periods of red soil development has been cited as the 
principal factor influencing evolution in land snail shells on Bermuda during the Quaternary. We argue that at no time 
was there an appreciable deficiency of calcium carbonate on Bermuda because the red soils themselves are largely made 
up of carbonate and because calcium in plant tissues would have been recycled during all stages of the Quaternary by 
frequent forest fires. Supposed instances of very localized changes in shell thickness arose through misinterpretation of 
chronology. Paedomorphic populations of the subgenus Poecilozonites occur only in carbonates of the last interglacial 
(Marine Isotope Stage 5) and the Holocene, never in glacial red soils as maintained repeatedly by Gould. 
INTRODUCTION 
In several studies, Gould (1966, 1968, 1969, 1970a, b, 
1971a, b) investigated evolutionary trends in diversity of 
shell size and shape in land snails, particularly of the 
genus Poecilozonites, subgenus Poecilozonites (Zoniti- 
dae), on the remote, oceanic island of Bermuda. He 
repeatedly hypothesized that the lack of calcium 
carbonate during glacial episodes, when snails were 
supposedly living on lime-poor red soils, was of prime 
importance in driving evolution in Bermudan snails. 
In attempting to explain repeated instances of 
paedomorphosis in the subgenus Poecilozonites, Gould 
(1968: 81) stated that: "The most paedomorphic 
subspecies originated in red soils; paedomorphs did 
not evolve in times of carbonate-dune deposition. The 
thin shells of paedomorphs might have been adaptive in 
the low-calcium environment of red soils." Gould 
(1977: 277) later unequivocally identified the "adaptive 
trigger" of paedomorphosis as "the almost totally lime- 
free soils that served as substrate for the most 
paedomorphic forms." 
Calcium-poor red soils supposedly not only ex- 
plained paedomorphosis but also influenced evolution 
in other lineages in the subgenus (Gould, 1969: 482- 
492). For example, in the giant species P. nelsoni "a 
thin-shelled subspecies hved in red soils" (Gould, 1969: 
491). 
In another study, Gould (1970a: 572) proposed that 
variations in relative abundance or shell morphology in 
six different species or populations of land snails on 
Bermuda during the last two glacial cycles "was 
influenced primarily by the availability of calcium 
carbonate for shell construction." 
And in yet another case (Gould, 1971a: 91), what 
were interpreted as very restricted local populations of 
several different snails were thought to differ from their 
supposed contemporaries elsewhere in the island 
because of local soil conditions, thus providing support 
for Gould's "previous assertion (Gould, 1968 and 
[1969], pp. 482^83) that the adaptive significance of 
paedomorphosis ... lies in the thin shell that it produces 
and that lime-poor habitats require." 
A much more refined knowledge of the stratigraphy 
and geochronology of Bermuda (e.g.. Vacher et al., 
1989, 1995; Hearty, 2002; Hearty et al., 2004) than was 
available to Gould presents a very different picture of 
the evolution of the island's snails. The progress 
achieved in unraveling the complex limestone architec- 
ture of Bermuda's sedimentary units and their ages can 
be partially attributed to traditional and new appUca- 
tions of amino acid racemization (AAR) geochronol- 
ogy (Hearty et al., 1992). We have determined the 
degree of epimerization of D-alloisoleucine/L-isoleu- 
cine (or A/I) from several populations of Poecilozonites 
including P. bermudensis bermudensis, P. b. fasolti, and 
other paedomorphs (Tables 1 and 2). AAR geochro- 
nology provides an independent means with which to 
determine relative ages and age-succession of the land 
snails in question. A/I is determined on several 
individual snails from each sample. A/I values are 
approximately 0.015 in living specimens, 0.40 to 0.58 
s. L. Olson & P. J. Hearty, 2006 Page 133 
Table 1 
Taxa of Poecilozoniies {Poecilozonites) discussed in this paper with their old and revised chronologies. Each of the 
main lineages {P. b. bermudensis, P. b. zonatus, and P. nebont) was considered by Gould (1969) to be continuous 
across their age ranges. The stratigraphie names used in Gould (1969) are poorly defined in terms of both stratigrapic 
position and apparent age (see Hearty, 2002). We offer only inferences of what Gould may have meant by use of 
these stratigraphie names. 
Old chronology (ages 
Taxon inferred from Gould, 1969)* 
Revised chronology 
(estimated age ka) 
This study 
Correlated 
marine isotope 
stage (MIS) Comments 
P. b, bermudensis 
(Pfeiffer, 1845) 
P. b. zonatus 
(Verrill, 1902) 
P. b. fasolii 
(Gould, 1969) 
P. b. siegmimdi 
(Site 6; Ireland Isl.) 
(Gould, 1969) 
P. b. siegUndae 
(Site 5; Rocky Bay) 
(Gould, 1969) 
P. nelsoni 
(Bland. 1875) 
St. Georges, Southampton,       Holocene (12 ka -present) MIS 1 only 
Recent (80 ka to present?) 
Shore Hills (>300 ka?) to St.   Interglacial intra-dune AU MIS 5 
Georges (80-10 ka?) and post highstand only 
soils (130-80 ka) 
Shore Hills (>300 ka?) Holocene cave coUuvium Late MIS 1 
(1-2 ka) 
Harrington red soil 
(post 125 ka?) 
Interglacial intra-dune soils      Mid MIS 5 
(c. 100-115 ka) 
Pembroke and Harrington       Interglacial intra-dune soils      Mid MIS 5 
red soil (post 125 ka?) (c. 100-115 ka) 
Shore Hills (>300 ka?) to St.   Glacial red soils (180-130 ka   MIS 6, 4-2 
Georges (80-10 ka?) and 80-10 ka 
The modern phytetic 
lineage evolved 
independently of 
paedomorphs in the 
Pleistocene. 
Considered to be the trunk 
species of aU P. 
bermudensis 
Differs only slightly from 
other late Holocene 
morphotypes 
AAR ratios indicate P.b. 
siegmundi and P.b. 
siegUndae are the same 
mid MIS 5 age 
Two additional unnamed 
last interglacial age 
paedomorphs are of 
similar (St. Georges), or 
slightly older ages 
(Hamilton) compared to 
the named subspecies 
Nearly identical P. nelsoni 
morphotypes are found 
in three successive 
glacial-age soils 
from shells from Marine Isotope Stage (MIS) 5 (last 
interglacial), and progress to an equilibrium ratio of 
1.30 after perhaps 1,000,000 yr (Hearty et al., 1992). 
With independent ^"C or U/Th age calibration at key 
intervals in the epimerization reaction, it is possible to 
compute absolute age estimates directly from A/I. 
Relevant discussion and protocols regarding this 
application and underlying database are available in 
Hearty et al. (2004). 
Before detailing evolutionary sequences and the 
possible causes in subsequent papers, we wish to 
dispose of the fallacious assumptions and erroneous 
facts associated with the hypothesis of the effects of 
carbonate limitation on the evolution of Bermudan 
snails. It is highly unlikely that any significant area of 
Bermuda was deficient in calcium at any period during 
the entire post-volcanic history of the island. In one 
lineage, snails during glacial periods deposited far more 
calcium in their shells than during interglacial times. 
Gould's (1971a) instance of highly local enviroimiental 
control was based on an erroneous assumption of the 
age of the sample. And finally, in complete contrast to 
Gould's repeated assertions, all of his fossil examples of 
paedomorphs existed during interglacial times of high 
carbonate deposition and never occur in glacial red 
soils, 
AVAILABILITY OF CALCIUM 
ON BERMUDA 
To begin with, Gould's correlation of shell variation 
with the amount of calcium carbonate in the substrate 
is compromised by Goodfriend's (1986) detailed review 
of causes of variation in shells of land snails. Gould 
(1968) cited investigations by Rensch (1932) and 
Oldham (1934) as showing calcium deprivation to 
result in shell thinning or mortality in snails. Yet the 
same papers are cited by Goodfriend (1986: 208) as 
finding ambiguous results for a correlation between 
shell size and limestone availability. Goodfriend (1986: 
208) concluded that field studies "failed to reveal any 
relationship   between   shell   size   and   the   calcium- 
Page 134 The Veliger, Vol. 49, No. 3 
Table 2 
Amino acid ratios on Poecilozoniles b. bermudensis ( Pbb), P. b. fasolli (Pbfas), P. b. siegUndae (Pbsl), P. b. siegmundi 
(Pbsm) and unnamed paedomorphs (paed) from Bermuda. Gould (1969) collections from the MCZ are identified by 
"SJG" prefix. Our collections from numbered Gould (1969) sites are preceded by our field number, then Gould's. 
NAU-AAL = sample number from Northern Arizona University Amino Acid Laboratory. Age interpretations not 
in years refer to Marine Isotope Stage (1 = Holocene; 5 = last interglacial). Gould's Tom Moore's Cave is thought to 
be the same as either Walsingham Cave or Walsingham Sink Cave. The date for Fern Sink Cave is based on 
a charcoal AMS '"C age of 1630 ? 60 yr BP (Hearty et al., 2004). 
S!te# 
Fig. 1 FMNH NAU-AAL FIELD # Spp Locality information Age/Interp. Mean St Dev N= 
? 4081 ANSP 85510 Pbb Alive 1903 104 yr 0.017 0.001 1 
2 303 J 92-3 3S41 SJG 5 3 Pbfas Tom Moore's Cave P. b. fasolli 
(Gould's collection) 
Late 1 0,055 0.034 2 
2 303194-5 3847 UGClz(2a) Pbb Fern Sink Cave upper 1,630 yr BP 0.047 0.014 3 
2 303196 4575-77 UWVl Pbb Walsingham Cave Late 1 0.045 0,017 3 
2 303197 4578-80 UWSl Pbb Walsingham Sink Cave, entrance Late 1 0.046 0,019 3 
2 303198 4403,05 UWSu2 Pbb Walsingham Sink Cave, deeper deposit Late I 0.069 0.008 2 
2 303199 3845 UGClx(2) Pbb Fern Sink Cave lower level Mid 1 0.097 0.023 3 
7 303200-1 4607-10 UTBl/2 Paed Tobacco Bay, St. Georges Late 5 0.40 0.03 3 
5 303202-4 790514.1 URB-SJG 44 Pbsl Rocky Bay, type P. b. siegUndae Mid 5 0.51 0.01 2 
6 303205-S 3838-29 um- SJGIO Pbsm Ireland Island, type P. h. siegmundi Late 5e 0.523 0.027 6 
8 303209 3837 USLld Paed Shell Depot Late 5e 0.542 0.05 3 
9 303210 3864 USSlc Paed Saltus School, 1989 Mid 5e 0.587 0.021 3 
carbonate content of the substrate" and that "exper- 
imental studies give similarly inconsistent results." 
But it is one thing to suggest that lime-poor soils 
drive snail evolution and another to show that snails on 
Bermuda ever experienced lime-poor soils. The island 
of Bermuda (Figure 1) is composed almost entirely 
(over 95% of surface rocks) of lithified dunes of 
calcium carbonate sand formed from skeletal remains 
of marine invertebrates and coralline algae. Dune 
deposition takes place during interglacial periods when 
sea levels are elevated (Bretz, 1960). During the 
depressed sea levels of glacial periods, when the entire 
Figure 1.    Location map of sites discussed in the text. 
Bermuda platform is exposed, marine carbonate supply 
is cut off and red soils form from diagenesis of 
limestone and from wind-borne dust (Bricker & 
Mackenzie, 1970; Muhs et al., 1990), which is more 
abundant in the atmosphere and polar ice during 
glacial periods (Glacuum &. Prospero, 1980). 
Because they are developed directly on limestone, 
most soils on Bermuda contain large amounts of 
calcitim carbonate (CaCOj), often in the form of 
carbonate sand or silt. Of the 45 samples in 9 soil 
profiles analyzed by Ruhe et al. (1961: their tables 3, 4), 
41 were more than 80% CaCOj, and only 4 were below 
14.6% (all from subsurface horizons in the more 
ancient soils). The upper and surface samples in each 
profile were all above 63% CaCOj. All levels in the 
deep deposits of Admirals Cave, which span the past 
120 kya and thus include both glacial and interglacial 
slope wash sediments, contain at least 70% CaCOs 
(Hearty et al., 2004). 
The thickest soil development on Bermuda occurred 
during a prolonged period of depressed sea-levels from 
about 800 kya to 450 kya?the Big Red Soil of Hearty 
et al. (2004) and Olson et al. (2005). No specimens of 
land snails are known from this period and all of those 
studied by Gould are younger. Regardless, it is unlikely 
thai snails in Bermuda ever lived on soils that were 
substantially lacking in calcium. 
Even during glacial periods when red soils develop, 
areas of high relief and any erosional features would 
expose bare carbonates, particles of which would be 
carried by wind or water to places where carbonates are 
s. L. Olson & P. J. Hearty, 2006 Page 135 
covered by soil. Furthermore, snails need not derive 
calcium directly from the substrate, as they would also 
obtain substantial amounts of calcium from plant 
detritus. Calcium is necessary for the formation of 
plant cell walls, where once incorporated it remains 
inert. Thus it is necessary for new growth, so that 
plants themselves would not thrive on truly calcium 
deficient soils. If there actually were local areas where 
soils were deficient in calcium, these would nevertheless 
receive calcium from plant detritus blowing from areas 
higher in CaCOj and from CaCOj dust from the shore 
and erosional features exposing limestone. 
The roots of larger plants on Bermuda such as 
palmetto- {Sabal bermudiana) and particularly cedar 
(Juniperus bermudiana) would be able to penetrate soils 
to take up calcium directly from underlying carbonate 
deposits. Calcium sequestered in plant tissues such as 
dead leaves, bark, and wood, would be released in ash 
after burning- Bermuda experienced numerous natural 
fires probably throughout its history. We found 
charcoal, sometimes in great quantity, throughout the 
entire sequence in Admirals Cave (location in Figure 1) 
representing the last 120 ky of Bermuda's history 
(Hearty et al., 2004). We have also seen evidence of 
extensive natural burning in glacial soils of MIS 10 
(about 300 ? 30 ky old) exposed in Bierman's Quarry 
(location in Figure I; see Olson et al. 2005 for 
information on this site). Thus, there were fires capable 
of releasing calcium stored in plant tissues throughout 
the entire the fossil record that Gould studied. In short, 
it is difficult to imagine any situation in which 
a Bermudan snail would be stressed to obtain sufficient 
calcium for shell building. 
TOTAL SHELL CARBONATE IN GLACIAL 
VERSUS INTER GLACIAL SNAILS 
It was Gould's (1969: 487) contention that "snails from 
red soils (glacial periods) tend to reach larger maximal 
sizes, have thinner shells and be smaller at a whorl than 
samples from eolianites." Shell thinning was supposed 
to be an indication of decreased calcium resources. But 
in none of his writings on the subject did Gould provide 
any quemtitative measures of shell thickness or mass of 
Bermudan snails. 
If calcium carbonate were ever a limiting factor in 
snail evolution in Bermuda we woiild expect this to be 
reflected in the total shell mass, regardless of thickness. 
From our excavations in the finely stratified deposits in 
Admirals Cave and from dating o? specimens from 
many other sites around the island, we can now be 
quite confident about the sequence of shell forms in the 
subgenus Poecihzonites since the penultimate glacia- 
tion (Hearty et al., 2004). The last two glacial periods 
are characterized by the presence of the giant form 
known as P.  nelsoni.  Although this form evolved 
independently in each of these glacial episodes, the 
forms are identical in size and shape so far as we have 
been able to determine. During the last interglacial, the 
widespread and abundant P. b. zonatus prevailed, and 
P. b. bermudensis characterized the present, Holocene, 
interglacial, although it is now perhaps extinct (Bieler 
& Slapcinsky, 2000). A sample of 10 shells of P. nelsoni 
had a mass of 37.7 g, whereas 10 shells of P. b. zonatus 
weighed 10.2 g, and 10 shells of fully mature individ- 
uals of P. b. bermudensis weighed 7.4 g, whereas 10 
from a lot of smaller shells weighed 5.0 g. Although 
these are pretty simple statistics, additional data are not 
going to alter the fact that in this lineage of 
Poecilozonites individual snails were depositing 4 to 5 
times as much calcium in their shells during glacial 
periods as their successors did during interglacials. This 
does not support Gould's idea of a deficiency of 
calcium carbonate during glacial periods. 
A CASE OF "UNUSUAL PRECISION" OF 
ENVIRONMENTAL CONTROL NEGATED 
Gould (1971a) cited a sample containing five species of 
snails from what he believed to be a Pleistocene red soil 
deposit (his "Shore Hills Soil") at a site that he called 
Tom Moore's Cave as being an example of an 
"extremely local event" in which very tight environ- 
mental control was exerted on shell morphology. In all 
five species the umbilical widths were greater than in 
"contemporaneous" samples from elsewhere on the 
island, meaning those that were also believed to be 
derived from the Shore Hills Soil. 
The large umbilicus was said to be negatively 
correlated with shell thickness (i.e., shells were thinner), 
which in turn was believed to be a response to low 
CaCOj levels in the soil. Thus, supposedly, over a small 
area of lime-deprived soil, all five species developed 
thin shells and large umbilici, whereas their supposedly 
contemporaneous nearby neighbors with greater avail- 
ability of lime did not. What Gould did not mention, 
however, is that the snails from Tom Moore's Cave are 
also exactly like the modem representatives of the same 
species. 
The greatest error to which Gould was subject in his 
studies of Bermudan snails was his belief that all 
samples from caves and fissures, which were associated 
with red soils, came from what he regarded as a middle 
Pleistocene Shore HUls Soil (Sayles, 1931) and were 
broadly contemporaneous. We now know, however, 
that snails and their enclosing soils in fissures and caves 
are not all contemporaneous. This is explained by the 
sedimentary process of fissure and cave filling. As 
a "pitfall" void opens to the surface, the capping soils 
on the host limestone, which may be hundreds of 
thousands of years old depending on the formation, 
would fill the void. In addition, organisms living at the 
Page 136 The Vchger, Yol 49, No. 3 
Table 3 
Simple inorphoratitrii; data from Bermuda Poccihzomtes paedomorphs discussed in (his study. Rows in bold are from 
Gould's MCZ collections, and in plain type from our (O&H) field toUections of Gould's and other relevant sites. 
Morphoinctrie measures of Gould's (1969) P. b, fasolti (SJC;53) and our sites from Walsiiigham Cave (UWVl) and 
Walsingham Sink Cave (UWSl) show no significant difi'erences. With the exception of UCGlx, P. b. fasolti and our 
samples are statistically identical in size and shape. On the basis of morphometric and amino a;;id data, we are 
confident that wc are dealing with the same populations as de.scribed in Gould (1969) and that P. b. fasolsi is Jate 
Holocene and the same as P. b. bernjutiensLs. 
Gould (1969) samples Shell width Shell height Total size 
our co?ections (O&It) (mm) lie (mm) ?1(7 Heighl/Width -lo (w J- h. mm) .Mo N 
SJG53P. b.fasoM 2J.76 1.26 S.Sl 0.38 t}.40 0.02 M>M 1.56 5 
CO&H) P. b. fusoM 21.82 1.26 8.53 0.62 0,39 0.?2 30.35 1.78 40 
(O&H) tlWVl Pbb 21.86 O.Sl S.6? 0.66 0.39 0.03 30.46 1,28 15 
(O&H) UWSl Pbb 21.91 1.73 8,49 1.08 0.39 0.03 30,41 2.71 15 
?O&H) UGClz (1) Pbb 21.76 2,43 S.76 1.62 0.40 0.03 30.51 4,00 7 
(O&H) UGClx (2) Pbb 19.15 1.40 7.43 0.87 0.39 0.02 26.58 2.20 6 
P, b, hernuideniis 20.05 5.44 7.40 1,27 0.37 o.w 27,45 6.72 2 
fO&H) Hoioceae Pbb 19.33 1.67 7.30 2.09 0.40 0.03 27.04 :.84 175 
(O&H) Pleistocene "Pbh" 19.52 1.28 7.68 o.es 0.39 0.02 27,19 ;.S2 30 
All pHedoiiiarphs 22.03 0.81 8.13 0.40 0.37 0.02 30.17 0.85 3 
P, i. Sffgfrmndi (Pbsm) 2i.m 8.50 039 J0,40 1 
fO&H) Phs,n 21.40 1.03 9.67 ?.?O 0.45 O.02 31.07 1.46 6 
CO&H) Pbsm 20.93 2.57 S.53 2.03 0.40 0.05 29.47 4,60 3 
P. b. s?gamiae (Pbsi) 21.30 7.70 0.36 2?.00 1 
(O&Hl Pbsl 21.SO 7.60 0.35 29.40 1 
surface that succumb to the same pitfaU would also be 
mixed with the ancient soils. This diachrony, explained 
in Hearty et al, (2004), was clearly not understood by 
Gould, leading to grsive misinterpretation of the 
biostratigraphic succession of Poedhzonites. Snails 
that Gould assumed were from his Shore Hills Soil, 
which he thought to be between 0,3 to 1 million yr old 
(MIS 9 to -25), include specimens from the J?LS? two 
glacial episodes (MIS 6 and 4/2), the iinerveniug 
interglacial (MIS 5), and the Holocene (MIS 1) (Heaity 
et al., 2004)! 
Wc have not been able to determine with absolute 
?nainty which of several caves near Tom Moore's 
Tavern is the one from which Gould obtained his 
sample, but from oui reconnaissance of the area it is 
most likely that this cave was one of two now known as 
Walsingham Cave and Walsingham Sink Cave (Olson 
et al., 2005), probably the latter. In Walsingham Cave 
the fossils in sediments that we found were entirely 
Holocene in age. In Walsingham Sink Cave, both in 
a rocksheller at the entrance and in a more extensive 
deposit deep in the cave, a thin veneer of loose Holocene 
sediments (by the inclusion of P. b. bermudensin) rested 
unconformably on a deeper, more compacted soil 
deposited during the last (Wisconsinan MIS 2-4) 
glacial stage, as determined from the inclusion of P. 
nelsoni and the flightless rail Rallus recessus (Olson & 
Wingate, 2001: Hearty et al., 2004; Olson et al., 2005), 
Regardless of the precise identity of Gould's Tom 
Moore's Cave, we have measured and analj-zed 
a sample of Gould's P. bermudensis fasotii from his 
own collection (SJG site #53; Gould, 1969: p. 507) at 
the MCZ (Tables 2 and 3), Two shells of P.b. fa.wlii 
yielded a mean AAR ratio of ?.055 ? 0.034. Although 
we have no independent '"C ages for these spcchnens, 
a single P. b. bermndenxis shell from nearby Fern Sink 
Cave (UGClz) produced a mean ratio of 0.047 i 0.014 
(3) with an AMS '"C age of 1.630 i -30 yr BP (Table 2). 
The estimated age of Gould's sample of P. b. fasolti, 
based on the age of the Fern Sink sample (Hearty et al., 
2?<M; and Figure 2), is 1900 ? ?200yr. The late 
Holocene age of the Tom Moore's Cave sample is 
unquestionable; the larger error being the result of the 
hmited number of shells analyzed. Additional samples 
from Walsingham Cave (UWVl) and Walsingham 
Sink (LTWSl) caves gave similar mean A/I values and 
morphometry (Tables 2 and 3), At this very rapid early 
stage of the epimerization reaction, small differences in 
age (i.e., several hundred years) would result in 
significantly difTerent A/J ratio."!. BecaUiiC the method 
is destructive wc chose to limit the number of shells 
damaged by sample extraction. 
To summarize, the sample of snuiU that Gould 
(1971a) attempted to usq to ?how "unusual precision" 
of environmental control in the middle Pleistocene is 
actually late Holoceno in ?tfc. only 2 ? 1 ky old, and is 
s. L. Olson & P. J. Hearty, 2006 Page 137 
20 
Age(xlOOOyr) 
Figure 2. Accelerator mass spectroinetry (AMS) '''C calibra- 
tion of Poecilozonites epimerization history over the past 
40 ka (Hearty et al., 20O4). Ail samples older than 10 ka are P. 
nelsoni, which has demonstrated a slightly faster (1.2X) 
epimerization rate than P. bermudensis, or P. b. zonatus. 
thus essentially modem in all respects. It tells us no 
more about evolution than any other modern sample. 
THE SOILS ON WHICH 
PAEDOMORPHS ORIGINATED 
Gould (1969) interpreted several populations of P. 
bermudensis, including the modem nominate subspe- 
cies, as paedomorphic offshoots that evolved indepen- 
dently at least 4 different times from non-paedomor- 
phic stock. We agree with some aspects of this 
interpretation, namely that the fonns are paedomorphs 
and that they evolved independently more than once. 
We will deal with the evolution of paedomorphs in 
another paper but here intend to take a closer look at 
Gould's extraordinary statement that the paedomorphs 
"originated in red soils; paedomorphs did not evolve in 
times of carbonate-dune deposition" (Gould, 1968: 81). 
The modem taxon Poecilozonites bermudensis ber- 
mudensis was Gould's prime example of a paedomorph. 
This form occurred throughout the island in historic 
times, although it was apparently extirpated in the 20th 
century by introduced predators and other human- 
induced enviroimiental factors. We now know that P. 
b. bermudensis evolved very rapidly at the beginning of 
the Holocene (Hearty et al., 2004), so it never existed 
in a glacial environment, or during a glacial episode. 
The Holocene is an interglacial episode with high 
concentrations of carbonates at the surface, although 
few of these siu'face carbonates are of Holocene age. 
Gould (1969: 480, fig. 20) thought that his P. b. 
fasolti was the earhest of the various paedomorphic 
offshoots, supposedly originating in the red middle 
Pleistocene "Shore Hills Soil." As discussed above, 
however, topotypical paratypes of P. b. fasolti (SJG 
#53) are only 2 ? 1 ky old, hence essentially modem. 
Thus   P.   b. fasolti becomes  a  synonym   of P.   b. 
bermudensis and is not associated with a glacial period 
or glacial soUs. 
We collected new material from most of the other 
sites in which Gould identified paedomorphs as well as 
in several additional sites that we discovered. We 
determined A^I ratios from this new material and also 
from Gould's original material stored at MCZ. As 
detailed below, all of these paedomorphs occurred in 
sand or protosols of various stages of the last 
interglacial (MIS 5). 
Two other paedomorphic forms that Gould (1969) 
named, P. b. siegmundi and P. b. sieglindae, were both 
considered by Gould to come from what was then 
called the Harrington Formation, which in Gould's 
own sequence (1969: 414, tab. 1) was attributed to the 
last interglacial period, thus directly contradicting his 
assertion that paedomorphs originated in glacial red 
soils. We collected topotypes of P. b. siegmundi on 
Ireland Island (Gould's site 10) from what we agree is 
a post-peak last interglacial (MIS 5e) highstand unit. 
These occurred in a white structureless, nearly pure 
carbonate 'protosol' (Vacher & Hearty, 1989) such as 
form during interglacials during pauses in eolian 
deposition when vegetation colonized dunes, and soils 
began to form. The incipient soil development ceased 
when renewed eolian activity buried the horizon. Mean 
A/I ratios of 0.52 ? 0.03 (n = 6) at Ireland Island 
confirm that deposition followed shortly after the peak 
MIS 5e highstand, not during a glacial episode and 
with no red soil. 
Similarly, the type locality of P. b. sieglindae at 
Rocky Bay (Gould's site 44) is a protosol immediately 
overlying Devonshire marine deposits. A/I ratios of 
0.51 ? 0.01 (n = 2) are equivalent to those of the 
Ireland Island paedomorphs, placing this taxon in the 
MIS 5 interglacial. Once again, no glacial period, no 
red soil. 
Paedomorphs occur in MIS 5 interglacial deposits at 
several other sites as well. Samples that Gould (1969: 
515) identified as P. b. bermudensis from his sites 11, 12, 
15, 16, 21, 24, and 27 probably all fall in this category. 
These must therefore have evolved independently of P. 
b. bermudensis, which did not originate until the 
Holocene (Hearty et al., 2004). We have found two 
new sites near Saltus School in Hamilton and at the 
Shell fuel depot in St. Georges (probably the same rock 
unit as Gould's site 11) that yielded A/I mean values of 
0,59 ? 0.02 (n = 3) and 0.54 ? 0.05 (n = 3), 
respectively. These values would place the paedo- 
morphs at, or shortly after, the peak of the last 
interglacial (MIS 5e). Our sites increase the number of 
last interglacial deposits yielding paedomorphs, and 
extend their geographic range to central and eastern 
locaUties in Bermuda (Figure I). Paedomorphs at 
Tobacco Bay (Gould site 27), also in St. Georges, 
occur in a protosol, but in this case are associated with 
Page 138 The Veliger, Vol. 49, No. 3 
late MIS 5 interglacial Southampton eolianites, thus 
being somewhat younger than the others mentioned 
above. ATI from the site is 0.40 ? 0.03 (n = 3), which is 
correlated with the peak of the Southampton marine 
transgression and associated highstand sediment flux 
on to the island (Vacher & Hearty, 1989; Hearty et al., 
1992). 
In summary, all Pleistocene paedomorphs of the 
subgenus Poecilosonites occur in nearly pure carbonate 
(>95%) soils in deposits of the last interglacial (MIS 5). 
Those from the current interglacial (Holocene) occur in 
siwface and pitfall deposits with high carbonate content 
(>70%). Whatever the cause may be for their 
paedomorphosis, it was not a result of living on "the 
almost totally lime-free soils" of glacial episodes 
(Gould 1977: 277). 
CONCLUSION 
A revised and much more accurate and detailed 
chronology of the geological deposits in which fossils 
occur now presents a very different picture of 
evolutionary sequences and events in the history of 
snails on Bermuda from anything envisioned by Gould. 
Changes in shell size and shape through the Pleistocene 
were rapid, repeated, and dramatic. While this new 
context for understanding change in Bermudan land 
snails should prove more instructive of evolutionary 
patterns than before, the underlining causes may be 
more difficult to discern. Calcium limitation was the 
only cause that Gould advanced and it became for him, 
to use one of his pet phrases, a deus ex machina. We 
believe that lack of sufficient calcium was probably the 
last problem any snail on Bermuda would ever have to 
face, so that other explanations will be required. 
Acknowledgments. We are indebted to Fred Collier for 
extending every courtesy during our study of collections at 
the Museum of Comparative Zoology, Harvard University, 
Cambridge, Massachusetts (MCZ). We thank R?diger Bieler 
for valuable comments on the manuscript and he and Jochen 
Gerber for providing curatorial support for our collections at 
the Field Museum of Natural History, Chicago (FMNH). We 
are also grateful to Gary Rosenberg for permitting access to 
specimens in the collections of the Academy of Natural 
Sciences of Philadelphia (ANSP). During many field trips to 
Bermuda we have greatly benefited from the assistance and 
cooperation of Wolfgang Sterrer and Lisa Green, Bermuda 
Aquarium, Museum and Zoo (BAMZ). Amino acid analyses 
were performed at Northern Arizona University Amino Acid 
Laboratory (NAU-AAL; D. Kaufman, Director). This is 
contribution 119, Bermuda Biodiversity Project, BAMZ. 
LITERATURE CITED 
BIELER, R. & J. SLAPCINSKY. 2000. A case study for the 
development of an island fauna: recent terrestrial 
mollusks of Bermuda. Nemouria 44:1-100. 
BLAND, T. 1875. Notes on certain terrestrial mollusks, with 
descriptions of new species. Annals of the Lyceum of 
Natural History of New York 11:72-87. 
BRETZ, J. H. 1960. Bermuda: a partially drowned late mature 
Pleistocene karst- Bulletin of the Geological Society of 
America 71:1729-1754. 
BRICKER, O. P. & F, T. MACKENZIE. 1970. Limestones and 
red soils of Bermuda, discussion. Geological Society of 
America Bulletin 81:2523-2524. 
GLACCUM, R. A. & J. M. PROSPERO. 1980. Saharan aerosols 
over the tropical North Atlantic?mineralogy. Marine 
Geology 37:295-321. 
GooDFRiEND, G. A. 1986. Variation in land-snail shell form 
and si2e and its causes: a review. Systematic Zoology 35: 
204-223. 
GOULD, S. J. 1966. AUometry in Pleistocene land snails from 
Bermuda: the influence of size upon shape. Journal of 
Paleontology 40:1131-1141. 
GOULD, S. J. 1968. Ontogeny and the explanation of form: an 
allometric analysis. Journal of Paleontology Memoir 2: 
81-98. 
GOULD, S. J. 1969. An evolutionary microcosm: pleistocene 
and Recent history of the land snail P. (Poecilozonites) in 
Bermuda.   Bulletin   of  the   Museum   of  Comparative 
Zoology 138(7):407-531. 
GOULD,  S.  J.   1970a.  Coincidence of climatic and faunal 
fluctuations in Pleistocene Bermuda. Science 168:572-573. 
GOULD, S. J. 1970b. Land snail communities and Pleistocene 
climates in Bermuda: a multivariate analysis of micro- 
gastropod diversity. Pp. 486-521 in E. L. Yochelson (ed). 
Proceedings   of  the   North   American   Paleontological 
Convention. Allen Press: Lawrence, Kansas. 
GOULD, S. J. 1971a. Environmental control of form in land 
snails a case of unusual precision. The Nautilus 84:86-93. 
GOULD, S. J. 1971b. Precise but fortuitous convergence in 
Pleistocene   land   snails   from   Bermuda.   Journal   of 
Paleontology 45:409-^18. 
GOULD,  S.  J.   1977.   Ontogeny  and  Phylogeny.   Harvard 
University Press: Cambridge, Massachusetts. 
HEARTY,  P.  J.   2002.  A  revision  of the late Pleistocene 
stratigraphy of Bermuda. Sedimentary Geology 153(1-2): 
1-21. 
HEARTY, P. J., H. L. VACHER & R. M. MITTERER. 1992. Age 
and   aminostratigraphy   of   Pleistocene   limestones   in 
Bermuda. Geological Society of America Bulletin 104(4): 
471-480. 
HEARTY, P. J., S. L. OLSON, D. S. KAUFMAN, R. L. EDWARDS 
& H. CHENG. 2004. Stratigraphy and geochronology of 
pitfall accumulations in caves and fissures,  Bermuda. 
Quaternary Science Reviews 23:1151-1171. 
OLDHAM, C. 1934. Further observations on the influence of 
lime on the shells of snails. Malacological Society of 
London Proceedings 21:131-138. 
OLSON, S. L. & D. B. WINGATE. 2001. A new species of large 
flightless rail of the Ralius longriostrislelegans complex 
(Aves: rallidae) from the late Pleistocene of Bermuda. 
Proceedings of the Biological Society of Washington 114: 
509-516. 
OLSON, S. L., D. B. WINGATE, P. J. HEARTY & F. V. GRADY. 
2005,  Prodromus of vertebrate paleontology and geo- 
chronology of Bermuda. Pp. 219-232 in J. A. Alcover & 
P, Dover (ods,). Insular vertebrate evolution:the palaeon- 
lologicul approach. Monografies de la Societat d'Hist?ria 
Natural de les Balears 12. 
MuHS, D.. C. A. BU.SH, K. C. STEWART. T. R. ROWLAND & 
R,   C   CRITTENDEN.   1990,   Geochemical  evidence  of 
s. L. Olson & P. J. Hearty, 2006 Page 139 
Saharan dust parent material for soils developed on 
Quaternary limestones of Caribbean and western Atlantic 
islands. Quaternary Research 33:157-177. 
PFEIFFER, L. 1845. Descriptions of new species of land shells 
from the collection of Hugh Cuming. Proceedings of the 
Zoological Society of London 1845:43-68. 
RENSCH, B. 1932. ?ber die Abh?ngigkeit der Grosse, des 
relativen Gewichtes und der Oberfl?chenstruktur der 
Landschneckenschalen von den Umweltsfaktoren. Zeits- 
chrift fur Morphologie und ?kologie der Tiere 25:757- 
807. 
RUHE, R. V., J. G, CADY & R. S. GOMEZ. 1961. Paleosols of 
Bermuda. Geological Society of America Bulletin 72: 
1121-1142. 
SAYLES, R. W. 1931. Bermuda during the Ice Age. American 
Academy of Arts and Sciences 66:381?468. 
VACHER, H. L. & P. J. HEARTY. 1989. History of stage-5 sea 
level in Bermuda; with new evidence of a rise to present 
sea level during substage 5a. Quaternary Science Reviews 
8:159-168. 
VACHER, H. L., M. ROWE & P. GARRETT. 1989. Geologic 
map of Bermuda. Oxford Cartographers: U.K., 1:25,000 
scale. 
VACHER, H. L., P. J. HEARTY & M. P. ROWE. 1995. 
Stratigraphy of Bermuda: nomenclature, concepts, and 
status of multiple systems of classification. Geological 
Society of America Special Paper 300:269-294. 
VERRILL, A. E. 1902. The Bermuda islands: their scenery, 
climate, productions, physiography, natural history and 
geology; with sketches of their early history and the 
changes due to man. Transactions of the Coimecticut 
Academy of Arts and Sciences 11:413-730. 
?>?: 
THE Mo UL- 
VELIGER 
A Quarterly published by 
CALIFORNIA MALACOZOOLOGICAL SOCIETY, INC 
Berkeley, California 
R. Stohier (1901-2000), Founding Editor 
ISSN 0042-3211 
Volume 49 October 1,2007 Number 3 
CONTENTS 
The Genus Paradons Bergh, 1884 (Nudibranchia: Discodorididae) in the Tropical Americas, 
and South Africa with the Description of a New Species 
YOLANDA CAMACHO-GARCI'A AND TERRENCE M. GOSLINER     105 
New Species of the Genus Caelatura Conrad, 1865 (MoUusca, Gastropoda, Barleeidae) from 
oifthe Brazilian Coast 
FRANKLIN NOEL DOS SANTOS AND RICARDO SILVA ABSAL?O     120 
Freshwater Mussel (Bivalvia: Unionidae) Causes Incidental Fish Mortality 
JAMES QAY) R, CORDEIRO     129 
Fallacies Underlying the Assumption of Calcium Limitation on the Evolution of Land Snails 
in Bermuda 
STORRS L. OLSON AND PAUL J. HEARTY     132 
Holoplanktonic MoUusca (Gastropoda) from the Gulf of Aqaba, Red Sea and Gulf of Aden 
(Late Holocenc-Recent) 
ARIE W. JANSSEN     140 
Taxonomy of the Family Neilonellidae (Bivalvia, Protobranchia): Miocene and Plio-Pleistocene 
Species o?Pseudoneiloneila Laghi, 1986 from Italy 
RAFAEL LA PERNA     196 
SHORT NOTE 
Sexual Dimorphism in Soft Body Weight in Adult Monetaria annulus (Family Cypraeidae) 
T IRIE AND B. ADAMS    209 
BOOK REVIEW 
Marine and Brackish Water Gastropoda of Russia and Adjacent Countries: an Illustrated 
Catalogue.    Yu. I. KANTOR, A. V. SYSOEV 
JAMES H. MCLEAN    212 
The Veliger (ISSN 0042-3211) is published quarterly in January, April, July, and October by the 
California Malacozoological Society, Inc., % Santa Barbara Museum of Natural History, 2559 
Puesta del Sol Road, Santa Barbara, CA 93105. Periodicals postage paid at Berkeley, CA and 
additional mailing offices. POSTMASTER: Send address changes to The Veliger, Santa Barbara 
Museum of Natural History, 2559 Puesta del Sol Road, Santa Barbara, CA 93105.