Page 11 - Korn_alii_2006

M. Korn et al. • Sister species within Triops cancriformis

Table 6 Results from the single factor analysis of variance (ANOVA) T. c. mauritanicus, specimens of the Moroccan haplotype
for morphological characters. The genetic haplotype group was groups exhibit the longest furcal spines, separating them from
considered as the fixed factor for each analysis (haplotype groups the ‘Portuguese’ (with the single exception of the ‘M 056’
were defined as subsets of sequences sharing diagnostic sites, without haplotype group; Appendix 2) and ‘Gitanilla’ (Extremadura)
consideration of singletons). The northern Spanish population was
populations, the latter having the shortest spines in this sub-
treated as a separate haplotype group due to clear differences in
morphological characters to specimens of the same haplotype but species. The southern Spanish haplotype has an intermediate
position, being clearly separated only from the ‘Gitanilla’
with a different mode of reproduction. P-levels shown are corrected
values, after application of sequential Bonferroni procedure. (and from the ‘M 058’) haplotype group (Fig. 5A; Appendix 2B).
North African T. c. simplex (haplotype groups ‘Tunisia’ and

Dependent variable for ANOVA model d.f. F P ‘T.c.s. M.’) and the northern Spanish population previously
assigned to this subspecies are almost identical in respect to
Telson length ratio 15 116.1 < 0.001
this character. The intermediate form has a mean value that
Apodous abdominal segments in females 15 6.2 < 0.001
No. of dorsal carina spines 9 33.2 < 0.001 is even lower than that of North African T. c. simplex. The
latter is reported to show a frequent tendency to a general

weakness in the strength of the armature (Longhurst 1955,
1958; in Notostraca, the armature consists of posterior
first a clade consisting of the Moroccan samples, and second, carapace spines, abdominal spines and telson spines, the
a clade formed of the Portuguese samples and the Spanish latter including the furcal spines).
samples from Gitanilla. The North African T. c. simplex sam-
ples cluster as a monophylum in an unresolved trichotomous Dorsal carina spines. Most T. c. mauritanicus investigated had
relationship with these two T. c. mauritanicus subclades. The numerous dorsal carina spines, often exceeding 50. However,
southern Spanish samples of T. c. mauritanicus from Extrem- three of the investigated populations of this subspecies
adura, Sevilla and Huelva group into a third clade that is the included specimens that had less than 20 spines, and counts
sister group of the rest of the lineage. as low as eight spines were observed, which is less than the
10 reported in the literature as being the maximum value for
Morphological re-analyses T. c. cancriformis (Table 3). In our samples, most specimens of
All three ANOVA models yielded significant results (P < 0.001; T. c. cancriformis showed 0–4 spines. However, the number of
Table 6). Thus, a Tukey post-hoc test was used to compute all spines ranged from 0 to 30 in the Austrian commercial kit
pairwise comparisons among populations included in each population, although most of these spines were extremely
model. small. This clearly exceeds the maximum value reported
hitherto and demonstrates that there is clear overlap in this
Telson length ratio. The ratio of furcal spine length to the dis- character among these subspecies. The discrepancy between
tance between the tip of the furcal spine and the anterior- our spine counts and the number of spines reported in the
lateral edge of the telson. High values of this ratio are literature was possibly caused by the fact that other authors
indicative of long furcal spines. Telson length ratio divides did not include the smallest spines in their counts. However,
the populations into two significantly different groups we found these spines to show a size gradient in most speci-
(ANOVA, P < 0.001; see Appendix 2B; Fig. 5A). It separates all mens, and their classification into ‘large’ and ‘small’ spines
T. c. mauritanicus (haplotype group means ranging from 0.38 would be arbitrary. In addition, Moroccan T. c. mauritanicus
to 0.50) from the remaining populations, which show much often showed a strong reduction in the size of carina spines,
lower values (haplotype group means: 0.21–0.27). Within attaining a condition more typical for T. c. cancriformis in one

Fig. 4 A, B. The hypotheses of Triops cancriformis phylogeny as reflected by our mitochondrial sequence data. —A. The first of two most
parsimonious trees (score = 129, CI = 0.8682, RI = 0.9742) based on the large 16S dataset and calculated with PAUP* (version 4.0b10, Swofford
1998; settings gapmode = new, add = cl). Bootstrap support is given above (or to the left of) selected branches calculated with PHYML
(nreps = 500; presented in percent)/ML-NJ (PAUP*; nreps = 1000)/MP (PAUP*; gapmode = new, maxtree = 1000, nreps = 1000). The 16S
haplotype groups (defined by diagnostic sites, see Appendix 1) are indicated to the right of each clade. —B. The PHYML reconstruction based
on the combined 12S and 16S sequences of a selection of samples and calculated with parameters corresponding to the TVM + G model
(evaluated by Modeltest 3.06; Posada & Crandall 1998): Base = (0.3553 0.1519 0.1515); Nst = 6; Rmat = (1.4243 9.7012 3.4980 0.0000001
9.7012); Rates = gamma; Shape = 0.2194; Pinvar = 0.0. Bootstrap support is presented in the same manner as in A. In both A and B the scale
enables comparison of evolutionary changes with the branch lengths, which are proportional to the evolutionary difference between taxa.
Selected nodes are labelled for convenient comparison of the two topologies. Abbreviations and symbols: T.c., Triops cancriformis; T.c.c.,
T. c. cancriformis incl. T. c. simplex from Girona (northern Spain); T.c.m., T. c. mauritanicus; T.c.s., North African T. c. simplex.

© 2006 The Authors. Journal compilation © 2006 The Norwegian Academy of Science and Letters • Zoologica Scripta, 35, 4, July 2006, pp301–322 311