geographic variation in allozymes of populations of salamandra

J. Evol. Biol. 9: 83- 102 (1996) IOIOL061X/96/010083 20 $ 1.50 +0.20/O

( 1996 Birkhiuser Verlag, Base1

Geographic variation in allozymes of populations of Salamandra salamandra (Amphibia: Urodela) exhibiting distinct reproductive modes

M. Alcobendas, H. Dopazo and P. Alberch*

Museo Nucional de Ciencia Nuturales (CSIC) C/ Jos4 Gutierrez Ahascul 2, E-28006 Madrid, Spuin

Key words: Urodeles; Salumundru sulumandra; Iberian Peninsula; genetic differentiation; microevolution; life history.

Abstract

The populations of the urodele Sulumundru salumundru in the Northern Iberian Peninsula exhibit very different coloration patterns and a remarkable range in reproductive modes (from giving birth to a large number of aquatic larvae to a parturition event of just a few fully metamorphosed, i.e. terrestrial, offspring). Electrophoretic study of geographic variation in allozymes shows that this extraor- dinary diversity, particularly in reproductive modes, is not accompanied by a genetic differentiation of similar magnitude. All the populations sampled along a transect crossing the Northern part of the Tberian Peninsula and encompassing the various reproductive strategies, as previously described, can be ascribed to a single species, because of small interpopulational genetic distances (ranging D,,,<~, from 0.05 to 0.199) and absence of fixed (diagnostic) alleles. A variety of phenetic and cladistic methods were used to elucidate the relationship among populations, based on allozyme data. These methods defined two well corroborated clusters: the first contains populations of salamanders with a blotched dorsal coloration pattern and characterized by parturition of aquatic larvae; the second group is composed of populations exhibiting a striped dorsal coloration pattern, smaller adult body size, and giving birth to fully metamorphosed terrestrial offspring. The latter group also encompasses some populations where mixed parturition events, which include both larvae and metamorphosed offspring, which have been recorded (Dopazo and Alberch, 1994). The absence of a correlation between genetic and geographic distance suggests that the mode of differentiation of the species is based on at least

* Author for correspondcncc.

83

x4 Alcobendas ct al.

two successive events of isolation, radiation, and secondary contact between populations. Furthermore, the validity of the described “subspecies” is questioned by our data, which point out the need for a detailed systematic study of Sulumandru from a global perspective. “Viviparity”, here meaning giving birth to fully metamorphosed offspring, originated once and occurs as intraspecific, and even as intrapopulational variation. Thus, we confirm a system where a major evolutionary innovation -the acquisition of independence from the aquatic media in the primitive amphibian complex life cycle-, can be studied at the microevolutionary, i.e., intra- and inter-populational level.

Introduction

The urodele Sulumundru salumundru is widely distributed throughout Europe, North Africa and the Middle East (Thorn, 1968). The species has undergone a major process of evolutionary radiation in the Tberian Peninsula, as evidenced by the fact that of the approximately fifteen geographical forms, described in the literature as “subspecies”, at least nine are found in the Iberian Peninsula, eight of them being endemic to this region (e.g. Eiselt, 1958; Thorn, 1968; Malkmus, 1983; Garcia-Paris, 1985; Klewen, 1991; Joger and Steinfarz, 1994).

S. .sufumandtz exhibits an amphibian complex life cycle, with the exception that the eggs hatch prior to, or simultaneously with, parturition. Thus, females go back to the water to give birth to a relatively large number of already hatched offspring (ordinarily ranging from thirty to sixty). The newly born exhibit a characteristic larval morphology with external gills as well as other physical traits associated with their aquatic lifestyle. After a variable period of time, ranging from a few weeks to several months or years (e.g. Joly, 196X; Bas, 1983; Alcobendas, Castanet and Alberch, submitted), the larvae undergo metamorphosis to become fully terrestrial (see Duellman and Trucb, 1986 for a review of larval morphology in urodeles and amphibian metamorphosis). This reproductive mode is usually referred to as “ovoviviparity” (Wake, 1982). Our usage of the terms, ovoviviparity and viviparity, follows the tradition in the Sulumand~u literature (e.g. Joly, 1968; Fachbach, 1976; Warburg et al., 1979; ijzeti, 1979). Such terms have a very specific meaning when applied to Sulumundvu, although they may not always be equivalent to the same terms when applied to other non amphibian taxa (Wake, 1982, 1993; Blackburn, 1994).

S. sulumundru spends its post-metamorphic life in the forest floor of woodlands. In the Northern Tberian Peninsula - our study area -, the habitat of S. sulumundt-u is fairly homogeneous in terms of climatic conditions and vegetation types: relatively high humidity and rainfall levels, abundance of rivers, brooks and permanent, as well as temporary, small water bodies in beechwood and oak forests (Bas, 1983; Garcia-Paris, 19X.5).

Wolterstorff ( 1928, subsequently expanded and elaborated by Gasser ( 1975) and Fachbach ( 1976)) described a subspecies, S. sulumund~u hernardelezi from a region in Northern Spain (Asturias and Cantabria), which has completely ehminatcd the

Allozyme variation in S. .sulumuntiru

Fig. 1. Dorsal coloration patterns and approximate geographic distribution of the four “subspecies” 01

SU/U~W&Y~ .~crlum~~t/r~ in the Northern Ibcrian Peninsula. A) General map of the Ibcrian Peninsula showing the 1500 meters level curve. The study area is indicated by a continuous line. B) Magnification of the study area with the approximate areas of distributions of the four geographical races (subspecies): l/S.s. pllricu: 2/S.s. hcmurrici; 3/S.s. fus/~o.v~; 4/S..s. fcrrestris. C) Patterns of coloration characteristic of each subspecies. Numbers as in B. “h~~~~rrrd~zi” and “/i~v~uo.vtr” are always striped in dorsal coloration and smaller in body siye than the other two forms which exhibit a dorsal coloration of more or less regular blotches.

aquatic larval stage (Fig. 1). S.S. hrvnm~/~~i females give birth to a few fully metamorphosed terrestrial offspring (from just one to a maximum of fifteen individuals per parturition event: Dopazo and Alberch, 1994). This unusual reproductive mode, for urodeles, is referred to as “viviparity” (e.g. Greven and Thies- meier, 1994).

In addition, Joly ( 1968) and Dopazo and Alberch ( 1994) have shown that populations ascribed to the subspecies S.S. firstuosa (Schreiber, 1912) exhibit inter-

Alcobendas et al.

Fig. 2. Summary of life history features characteristic of the four races of populations studied

and intrapopulational variability that includes both of the previously described reproductive modes. That is, some females give birth to larvae while others to metamorphosed offspring. Dopazo and Alberch ( 1994) cvcn report single parturition events involving both, metamorphosed and larval siblings.

Our study area, cxtcnding along the Northern side of the Tberian Peninsula, encompasses all of the outlined reproductive modes. Four races (subspecies) have been described from this area. Their morphology, characteristic patterns of dorsal coloration and approximate distribution are shown in Fig. 1. Although this article questions the evolutionary meaning of the described subspecies as monophylctic assemblages, we use the names as descriptors of morphology and geographic distribution. Besides the two reproductive modes (ovoviviparous and viviparous) defined above, coloration patterns sort the populations studied into two groups: S.S. &him and S.S. terrestris have a dorsal coloration of yellow blotches on a black background. S.S. gulluicu is characterized by sparsely distributed red blotches

Allozyme variation in S. .salamcmtirtr 87

in addition to the yellow markings. S.S. galluica and S.S. tervwtris are larger in body size as well as ovoviviparous in reproduction mode. The viviparous S.S. hrvnardezi and S.S. ,f&tuosa, with a mixed reproductive mode, are smaller in body size and striped in dorsal coloration (see Figs. 1 and 2 for a summary or morphological and life history information).

Although authors have assumed that the above subspecies were conspecific, there is no empirical evidence, besides morphological similarity, supporting the hypothesis of a single species. In fact, the striking differences in reproductive modes would question such an interpretation. Otherwise, the single species hypothesis constitutes a unique evolutionary system of intraspecific variability in life history patterns. A model system to address a fundamental issue in evolutionary theory: the origin of novelties that open up new adaptive realms (e.g. Nitecki, 1990). The elimination of the aquatic stage from the species life cycle clearly constitutes a “key innovation”. Hence, it becomes critical to adequately verify the status of the populations as a single biological species. We explore the patterns of geographic variability in allozymes in an attempt to reconstruct the evolutionary relationships between the populations exhibiting different reproductive modes. This preliminary biochemical survey was necessary given the few, and inadequate, available studies of the systematics of S. sulumundrra using molecular techniques (Fachbach, 1971; 1976; Gasser, 1975; 1978a). In addition, the results reported here have allowed us to identify areas of contact, and possible gene flow, between groups characterized by distinct life history patterns. This information provided the necessary background to carry out more detailed analysis of the patterns of gene flow in the contact zones using mtDNA haplotypes (Dopazo, Boto and Alberch, submitted) and DNA sequencing (Garcia-Paris, Alcobendas and Alberch, submitted).

Table I. Localities for Suhnandru sulwwxh. Numbers refer to populations in Figure 1. Subspecies

designated according to Gasser ( 1975).

Subspecies Number and Locality Altitude N

I-Arguenos (Hte Garonne Fra rice) 800 m. 2-Montseny (Cataluiia) IlOOm. 3-Pontcvcdra (Galicia) 200 m. 4-Fonsagrada (Galicia) 1000 m. 5-0~~0s (Galicia) 680 m. 6-Puerto del Palo (Asturias) 1100 m. 7-Godan ( Asturias) 240 m.

X-Covadonga (Asturias) 600 m. 9-Oviedo (Asturias) 200 m. IO-Mirador del Fito (Asturias) 600 m.

I I-Ucieda (Cantabria) 300 m. I Z-Law (Navarrd) 700 m. 13.Collado Lindus (Navarra) I I30 m.

(10) (10) (8)

(10) (10) (10) (8)

( 10) (13)

(8) (10)

(5) (10)

N = size sample.

88 Alcobcndas et al.

Table 2. Electrophoretic conditions for 34 proteins examined in thirteen populations of Snlmmunt/ra .XllWWUitYl.

Protein E.C. no. Locus ButTer Tissue Number of system extract alleles

Albumin

Aspartate aminotransferase

Aconitase hydrdtase Adenylate kinase

Creatine kinase Esterase

Fructo-Kinase Fumarase

Glucose dehydrogenase Glutamate dehydrogenase Glycerol-Sphosphate dehydrogenase Glucose phosphate isomerase Hcxokinase

L-lditol dehydrogenase Isocitrate dehydrogenase

Leucine Amine peptidase

Lactate dehydrogcnase

Malate dehydrogenase

Malate dehydrogenase NADP + -dependent Mannose phosphate isomerase Peptidase

Phosphogluconate dehydrogenase Phosphoglucomutase Pyruvate kinase Sorbitol dehydrogenasc Xanthine dehydrogenase

Alh-I LiOH P 2 AU-2 LiOH P 4

2.6.1.1 Ad-1 Poulik H,L 1

Auf-2 TC6.7/Poulik H,L 4 4.2.1.3 ANI TME6.9 H,L 2 2.7.4.3 Ak-I TC6.7 M 2

Ak-2 TC6.7 M 2

2.7.3.2 Ck TC6.7 M 2 3.1.1.1 E.st -11 TME6.9 H, L 3

Esf -m TME6.9 M 5 3.7.1.4 Fk TC6.7 H,L 1 4.2.1.2 Fum TME6.9 H,L I I. I. I .47 Gdh TC8 + H,L 5 I .4. I .2 Gld TCXt M 1 1.1.1.8 G3plh Poulik H,L 2

5.3. I .9 Gpi TC6.7/LiOH M 4 2.7.1.1 Hk-h TCX + II,L 3

Hk-m TC8 + M 1 1.15.1.1 Iddh LiOH H.L 3 1.1.1.42 I&l TC8 t H.L 2

Idh-2 TC8 + H,L 2 3.4.11 LUP TCX + H,L 3 1.1.1.27 Ldh-A TC6.7 H,L, M I

Ldh-B TC6.7 H. L, M 4 1.1.1.37 Mdh-1 TC6.7 H,L 1

Mdh-2 TC6.7 II, L 2 I. I. 1.40 Mdhp TC8 + H,L I 5.3.1.8 Mpi TC8 + H,L 2 3.4.11/13 Pel’ TC8 + M 3 1.1.1.44 P,qth TC8 + M 2 2.7.5.1 Pgm TC6.7 H.L 4 2.7.1.40 Pk TC6.7 M I

1 .I .I .I4 Sf//l TCX H,L 3 I, I, I .204 Xdh TC6.7 H,L I

Buffers for starch gel electrophoresis: TC6.7 = Tris-citrate pH 6.7 (I); TC8 = Tris-citrate pH 8.0 ( I); TME6.9 = Tris-Maleate-EDTA pH 6.9 ( I); LiOH = Lithium Hydroxide (I); Poulik = Tris-citrate borate

pH 8.7 ( I). (I) = Pasteur et al. (1987). Tissues: H = heart, L = liver, M = muscle, P = plasma.

Material and methods

Thirteen populations from the Northern part of the Iberian Peninsula were chosen along an east-west transect designed to include all the described subspecies in the region (Gasser, 1975; Fachbach, 1976; Gasser, 1987a) and to focus on the

Allozyme variation in S. .saltmcmh 89

transitional areas between the viviparous and ovoviviparous reproductive modes (Tab. 1; Figs. 1 and 2). Specimens were freshly dissected for blood, heart, liver and muscle samples. Blood samples were centrifuged at 1500 rpm, separating the plasma supernatant and the red cell fraction. The other tissues were homogenized in a buffer solution (0.1 M Tris, 10-3 EDTA, 5 x 10m5 NADP adjusted to pH 6.8 with HCl; see Pasteur et al., 1987) and centrifuged. The supernatant of the different tissues, the plasma and the red cell fractions were stored at -70 C until use. The specimens will be deposited in the collections of the Museo National de Ciencas Naturales (MNCN), Madrid.

Thirty four loci coding for proteins (including plasma proteins) were studied by means of standard horizontal starch-gel electrophoresis, using the histochemical staining procedures of Selander et al. (1971) Harris and Hopkinson (1976) and Pasteur et al. ( 1987) (see Tab. 2 for methodological details). Multiple loci arc designated numerically, the slowest migrating being the number 1; multiple alleles are designated alphabetically, the fastest (or most anodal) migrant being “a”.

Genetic variation among populations was calculated using the Biosys-1 software (Swofford and Selander, 1981). Modified Rogers genetic distances (Wright, 1978) were used to generate the Distance-Wagner unrooted tree (Farris, 1972). The Distance-Wagner unrooted tree does not assume a constant evolutionary rate of the proteins and consequently shows a more realistic representation of the relations between the Operational Taxonomic Units (OTUs) compared (Swofford, 1981). This procedure has been chosen in our study since Hillis (1985) and Shaffer et al. (1991) have shown it to be an adequate method to test the monophyly of the OTUs of a single taxon. Neighbour Joining (Phylip program; Felsenstein, 1989), was also applied to our data set to generate an unrooted tree of phenetic distances. Phylogenetic analysis was carried out using the Continuous Maximum Likelihood (CONTML; Felsenstein, 1989) procedure to estimate phylogcnies from gene frequency data (Felsenstein, 1989). The data were run ten times to increase the probability of finding the tree with greatest total likelihood. Following Shaffer et al. (1991) recommendation, to assess the robustness of the likelihood tree a jackknife test was applied to the data set. The results were summarised using the CON- SENSE program of Phylip. Because of the severe limitations of the CONTML procedure, estimates of phylogenies from distance matrix data under the additive tree model were carried out using the Fitch-Margoliash criterion (FITCH in Phylip programs). The programs were run ten times. The two procedures (CONTML and FITCH) do not assume an evolutionary clock.

The degree of correlation between genetic distance and geographic distance was explored by a Mantel test (MAXCOMP method in the NTSYS package; Rohlf, 1993). The populational level of genetic differentiation was explored using the standardized gene frequency variance F,, (we used BIOSYS-1 for its calculation) (Wright, 1965; Nei, 1977; Wright, 1978). This value ranges theoretically from zero (all genetic variance is partitioned exclusively within populations) to one (genetic variance is partitioned entirely among populations). The Fz’s parameter, or inbreeding coefficient, which ranges from - 1 (excess of heterozygotes) to 1 (deficit of heterozygotes), measures the deviation from random mating in the populations.

90 Alcobendas et al.

Fig. 3. Phenetic unrooted tree obtained by the Neighbor Joining method, as a representation of allo7yme

divergence. For each population, we report within brackets: the number assigned to the populations that correspond to the number in the map and the reproductive mode characteristic for the population. LAR.: birth of aquatic larvae; MET.: metamorphosed offspring: MIX.: mixed reproductive mode (larvae and metamorphosed). The coloration pattern of the populations is also diagrammatically shown next to

the corresponding branch.

Results

We identified 81 different electromorphs with the 34 loci examined among the thirteen populations studied here (Table 3). Ten of these 34 loci are monomorphic

Allozyme variation in S. sulu~rzcrnrlra 91

Fig. 4. Wagner unrooted tree obtained with Rogers’ modified genetic distance (Wright. 197X). Cophe-

netic correlation coelticient = 0.95. Population numbers as in Fig. 3.

(Ad-l, Fk, Fum, Gld, Hknl, L&-l, Mdh-1, MC, Pk, Xdh). An average of 1.46 alleles per locus is found, ranging from I .3 to 1.7 depending on the population.

Among the thirteen populations, there were no diagnostic alleles (i.e. fixed in all the individuals of just one of the populations; see Table 3). However, thirteen of the Xl alleles found in this study are present in just one of the populations, but generally in low frequencies. No exclusive alleles are found in the populations of Lanz, Covadonga, Palo and Oscos. Ucieda shows four exclusive alleles in four different loci (Ad-2, Gpi, Hkh, Ldh-2) Arguenos exhibits two of these alleles (Cd/r, Pgm) and the other seven populations just show one exclusive allele, at one locus.

Whichever method we used to analyse our data set, we always found the thirteen populations to be arranged into two distinct groups always including the same populations: the group A clustering Arguenos, Montseny, Pontevedra, Fonsagrada, Oscos and the group B composed of Lanz, Lindus, Ucieda, Oviedo, Fito, Cova-

Table

3.

Al

lele

fre

quen

cies

for

34

stud

ied

loci

in

th

irtee

n po

pula

tions

of

Su

lrm?a

ndra

su

irman

drcr

. Po

pula

tion

num

bers

co

rresp

ond

to

Tabl

e I.

E

Locu

s I

2 3

4 5

6 7

8 9

10

11

I2

I3

Alh-

1

Ah-2

Au/-l

Am

-2

Am

Ak-1

Ak-2

Ck

ESf

-II

ESZ-

WI

Fk

FW?l

GC

dh

a( 0

.89)

b(

O.l

I)

b a a(O.

lO)

c(O

.90)

a b a a b a a a a(O.

lO)

b(0.

25)

c(O

.55)

d(

0.10

)

a a(0.

11)

b(0.

89)

a a( 0

.20)

b(

0.4

5)

c( 0

.35)

a b a a b a a a b(0.

05)

c(O

.80)

d(

0.15

)

a(0.

80)

b( 0

.20)

a(

0.12

) b(

0.88

)

a b( 0

.75)

~(

0.25

)

a b a a a(0.

25)

b(0.

75)

a(0.

37)

b(0.

50)

c(O

.13)

a a c(O

.50)

d(

0.50

)

a b a b(0.

90)

c(O

.10)

a b a a(0.

95)

b(0.

05)

a(0.

56)

b( 0

.44)

a(0.

20)

b( 0

.80)

a a ~(0.

89)

d(0.

11)

a a(0.

05)

b(0.

95)

a b(0.

70)

c(O

.30)

a b a a a(0.

20)

b( 0

.70)

c(

O.1

0)

a(O.

10)

b( 0

.90)

a a c( 0

.40)

d(

0.60

)

a b(0.

20)

c(O

.80)

a b(0.

50)

c(O

.50)

a b a a a(O.

lO)

b(0.

90)

b a a b(0.

05)

c(O

.50)

d(

0.45

)

a c a b(0.

25)

c(O

.75)

a b a(0.

87)

b(0.

13)

a a(0.

12)

b(0.

88)

b( 0

.62)

c(

O.3

8)

a a c(O

.12)

d(

0.88

)

a C a b(0.

25)

c(O

.75)

a b a a a(O.

lO)

b(0.

90)

b(0.

90)

c(O

.10)

a a c(O

.70)

d(

0.30

)

a ~(0.

08)

d(0.

92)

a b(0.

35)

~(0.

65)

a a(0.

15)

b(0.

85)

a a(0.

38)

b(0.

62)

a(0.

23)

b(0.

57)

c(O

.20)

b(0.

31)

c( 0

.54)

d(

0.15

) a a b(

0.23

) ~(

0.42

) d(

0.35

)

a d a b(0.

25)

c(O

.75)

a b(0.

80)

a a(0.

80)

b(0.

20)

b(0.

75)

~(0.

25)

b(0.

85)

c(O

.15)

a a c(O

.40)

d(

0.60

)

a d a a(0.

05)

b(0.

20)

c( 0

.70)

d(

0.05

) a a(

0.20

) b a a a(

0.2

0)

b(0.

80)

b(0.

44)

~(0.

56)

a a ~(0.

80)

d(0.

20)

a ~(0.

63)

d(0.

33)

a a(0.

75)

~(0.

25)

a b a a(0.

60)

b( 0

.40)

b b(0.

60)

c(O

.40)

a a b(0.

12)

c(O

.12)

d(

0.38

) e(

0.38

)

a(0.

78)

b(0.

22)

b(0.

11)

c(O

.89)

a a(0.

85)

b(0.

05)

c(O

.10)

a(0.

90)

b(O.

lO)

a a( 0

.90)

b(

O.lO

)

b a(O.

15)

b(0.

75)

d(O.

lO)

a b

a k E

b(0.

20)

E c(

O.1

5)

d(0.

50)

2 d(

0.15

)

Gld

G3p

dh

Gpi

Hk-h

Hk-m

Id

dh

Idh

1

Idlz

-2

LUP

Ldh-

A Ld

h-B

Mdh

- 1

M

dh

-2

Mdh

p M

pi Pep

Pgdh

Pgm

Pk

Sdh

Xdh

b a b(0.

95)

c(O

.05)

b b a b(0.

90)

c(O

.10)

a a a b a(0.

80)

a(O.

20)

b(0.

80)

a( 0

.75)

c(O

.25)

a b(0.

15)

c a

a b c b a $%y

b a(0.

05)

b(0.

95)

a a b(0.

80)

c( 0

.20)

a a a b a( 0

.60)

c(

O.0

5)

a(0.

10)

b( 0

.90)

c i(O.1

5)

b(0.

31)

c(O

.70)

a

a b c b a b b b a a b(0.

94)

~(0.

06)

i(O.8

7)

b(0.

13)

a b a(0.

75)

b b(0.

12)

~(0.

88)

i(0.5

6)

b(0.

20)

c(O

.12)

a

a b C b i(O.7

0)

b( 0

.30)

b b a(0.

90)

b(0.

10)

a a(0.

05)

b( 0

.70)

c(

O.2

5)

a a a b c( 0

.65)

b c(O

.85)

d(

0.15

)

i(O.2

0)

$2:;

a

a b b(0.

10)

c(O

.90)

b a a(0.

05)

b(0.

90)

c(O

.05)

b b a(0.

67)

b(0.

33)

a i(O.9

5)

b(0.

05)

i(O.2

1)

b(0.

79)

c(O

.89)

b c( 0

.90)

d(

0.10

) a b(

0.2

2)

c(O

.70)

a

b a b a(0.

15)

b(0.

85)

a C t(O.8

0)

b(0.

20)

a a(0.

15)

b(0.

85)

a(0.

14)

~(0.

36)

b c( 0

.90)

d(

0.10

) a ~(

0.78

) a

b a b(0.

63)

c(O

.37)

b b a( 0

.62)

b(

0.36

)

a b(0.

31)

~(0.

69)

a a a a( 0

.43)

b(

0.57

)

c( 0

.50)

a(

0.12

) b(

0.8

8)

C a C a

i(O.4

0)

b(0.

60)

C b a b b(0.

20)

c(O

.80)

ii(O.9

0)

b(0.

10)

a b a(0.

44)

b c(O

.95)

d(

0.05

) a b(

0.08

) C a

a b C C a a i(O.0

4)

b(0.

96)

a(0.

38)

c( 0

.23)

a(

0.18

) b(

0.82

)

C a c( 0

.92)

a

i(0.6

5)

b(0.

35)

:yt82

oq)

b Et(O

.40)

$z$

a(0.

05)

b(0.

95)

b a(0.

25)

:g

,“,:’

a C a a a b a(0.

50)

b C a b(0.

10)

c a

a(0.

05)

b( 0

.95)

b a(0.

10)

$%]

a ~(0.

80)

d(0.

20)

a a a b a(0.

25)

~(0.

06)

a(0.

62)

b(0.

38)

C i(O.4

0)

c(O

.50)

a

i(O.8

7)

b(0.

13)

:gy

d(0.

10)

b b a(0.

37)

b(0.

50)

c(O

.12)

a C a a a b a(

0.2

5)

~(0.

62)

b b(0.

10)

c(O

.90)

i(O.7

5)

~(0.

25)

a

i(O.9

0)

b(0.

10)

b :(o.lo

) b(

0.60

) c(

O.3

0)

b a a a b a(0.

17)

~(0.

28)

a(0.

10)

b(0.

90)

c(O

.95)

d(

O.05

)

i(O.7

0)

c(O

.30)

a


Fig. 5. Phylogenetic unrooted trees obtained by way of: A) Continuous Maximum Likelihood procedure. Distances (in Km.) between the neighboring populations are superimposed. B) Fitch-Margoliash

method. Locality numbers as in Fig. 3.

donga, Palo and Godan. In the first group, we find an unexpectedly close relationship between populations from the Pyrennean mountains in the Eastern corner of the Peninsula and our most western locality in Galicia, at the extreme opposite of our range. In the second cluster, the populations from the central region of our study area (Asturias and Cantabria) are found to form a single cluster (Fig. 3). In each cluster, populations from different subspecies were pooled together: “terrestris” and “gallaica” in group A, “bernardezi” and “fastuosa” in group B.

The Neighbor Joining tree shows group A encompassing all the populations with strict ovoviviparity and more or less regular blotched pattern of coloration, while the populations with striped coloration pattern, and strict or optional viviparity, form group B (Fig. 3). The distance Wagner tree obtained with the modified Rogers genetic distance (Wright, 1978) identifies the same two clusters outlined above (Fig. 4).

Further phylogenetic analysis, based on the continuous Maximum Likelihood (CONTML) method and the Fitch-Margoliash procedure (FITCH), were also performed. The unrooted tree obtained with the CONTML procedure (Fig. 5a) shows very minor differences when compared with the distance Wagner tree (Fig. 4). Only the positions of Palo and Godan are reversed in relation to the Wagner tree. In the likelihood framework, the placement of the nodes cannot be unambigu-

. 2 Ta

ble

4.

Matr

ix of

ge

netic

di

stan

ce

coef

ficie

nts

by

com

paris

ons

of

thirt

een

popu

latio

ns

of

S.

salam

andr

a, an

d m

ean

hete

rozy

gosit

y fo

r ea

ch

5 po

pula

tions

(w

ith

stan

dard

de

viatio

n in

to

brac

ket).

Be

low

diag

onal

: M

odifie

d Ro

gers

di

stan

ce

(Wrig

ht.

1978

). Ab

ove

diag

onal

: Ne

i (1

978)

un

bias

ed

2.

gene

tic

dist

ance

.

z 6’

3

Popu

latio

n 1

2 3

4 5

6 I

8 9

5’

10

11

12

13

9 e 1

ARGU

ENOS

0.

034

0.07

2 0.

120

0.10

8 0.

161

0.13

8 0.

156

0.15

2 0.

165

0.18

3 0.

194

0.19

9 g

2 M

ONT

SENY

0.

187

0.02

5 0.

061

0.06

2 0.

121

0.11

5 0.

122

0.11

2 0.

139

0.16

2 0.

166

0.17

0 3

3 PO

NTEV

EDRA

0.

262

0.16

8 0.

049

0.04

8 0.

118

0.11

8 0.

132

0.12

5 0.

144

0.17

2 0.

155

0.16

2 s B

4 FO

NSAG

RADA

0.

327

0.24

3 0.

221

- 0.

029

0.09

6 0.

115

0.12

5 0.

117

0.13

1 0.

170

0.16

5 0.

177

c

5 os

cos

0.31

2 0.

244

0.21

9 0.

177

~ 0.

050

0.05

8 0.

099

0.09

2 0.

097

0.14

4 0.

118

0.13

3

6 PA

LO

0.37

1 0.

328

0.32

3 0.

295

0.22

2 0.

024

0.03

4 0.

077

0.05

8 0.

085

0.08

0 0.

072

7 GO

DAN

0.34

8 0.

322

0.32

5 0.

320

0.23

8 0.

170

- 0.

053

0.07

4 0.

079

0.10

5 0.

085

0.08

7

8 CO

VADO

NGA

0.36

7 0.

330

0.34

0 0.

332

0.29

9 0.

188

0.22

9 0.

071

0.05

7 0.

098

0.09

6 0.

089

9 OV

IEDO

0.

359

0.31

4 0.

328

0.31

8 0.

286

0.26

4 0.

262

0.25

6 0.

043

0.07

9 0.

105

0.15

3

10

FIT0

0.

376

0.34

8 0.

353

0.33

7 0.

295

0.23

6 0.

271

0.23

4 0.

207

0.04

5 0.

069

0.09

1

11

UCIE

DA

0.39

0 0.

371

0.37

9 0.

376

0.34

9 0.

277

0.30

5 0.

296

0.26

6 0.

214

0.07

3 0.

085

12

LANZ

0.

408

0.38

2 0.

370

0.38

0 0.

329

0.28

2 0.

289

0.30

4 0.

310

0.26

6 0.

271

0.00

5

13

LIND

US

0.40

5 0.

379

0.36

9 0.

383

0.33

8 0.

259

0.28

2 0.

285

0.35

4 0.

286

0.27

6 0.

138

Sam

ple

size

Hete

rozy

gosit

y

-Dire

ct-c

ount

-Hdy

Wbg

ex

pecte

d

10

10

8 10

10

10

8

10

13

10

IO

5 IO

0.05

6 0.

053

0.06

6 0.

62

0.06

9 0.

055

0.07

5 0.

058

0.12

3 0.

112

0.08

9 0.

099

0.09

7

(0.0

22)

(0.0

52)

(0.0

27)

(0.0

26)

(0.0

24)

(0.0

20)

(0.0

32)

(0.0

27)

(0.0

27)

(0.0

31)

(0.0

30)

(0.0

32)

(0.0

31)

0.09

1 0.

088

0.12

2 0.

112

0.12

9 0.

139

0.14

4 0.

110

0.17

2 0.

134

0.17

5 0.

179

0.16

7 (0

.029

) (0

.031

) (0

.034

) (0

.032

) (0

.032

) (0

.036

) (0

.036

) (0

.031

) (0

.042

) (0

.037

) (0

.039

) (0

.045

) (0

.038

)

96 Alcobendas ct al.

Fig. 6. Relation between geographic and genetic distances. The comparison is performed with MX- COMP, using a Mantel test. Normalised Mantel statistic: r = 0.25. Approximate Mantel f-test: I = I .c)l.

ously resolved with our data set since internal branches have confidence intervals that include 0. In the latter tree, two groupings within cluster B appear unresolved: the relation between Palo and Godan, and the group formed by Ucieda, Oviedo and Fito. The tree obtained with the Fitch-Margoliash method provided additional confirmation regarding the clustering of the populations in two groups (Fig. 5b). In group B, the populations which showed a new position in the Fitch-Margoliash tree were those which formed ambiguous nodes in the likelihood tree.

All the trees show that within each group, A and B, the populations dealing with different subspecies are not sorted out in a well defined fashion. For instance, in cluster B, Ucieda (S.S. fu.stuosa) is closer to Fito (S.S. hernardezi) than to Lindus (S.S. Jicstuosa). In cluster A, the Montseny population (S.S. terrestris) exhibits similar genetic distance to Arguenos (S.S. terrestris) than to Pontevedra (S.,s. gallaicu).

Genetic versus Geographic distance

Nei (1978) unbiased genetic distances (DN) and modified Rogers distances (D,; Wright, 1978) for all the inter-populational comparisons are shown in Table 4. The levels of protein differentiation among populations range between D, = 0.005 (Lanz-Lindus) and D, = 0.199 (Lindus-Arguenos). (The higher values of D,, compared with D, values, just reflect the modified equation for small sample size).

When genetic distance is compared with geographic distance, there is no direct relation between the two variables. The most evident example is between Arguenos

Allozyme variation in S. suhnmdru 97

and Lindus, which exhibits the highest genetic distance (D, = 0.199) and a relatively short geographic distance ( 160 Km). Conversely, Montseny and Pontevedra, separated by 887 km, are related by a minuscule genetic distance of 0.025. The comparison of the two data sets, genetic distance and geographic distance using the MAXCOMP method, indicates that these two parameters are not correlated (r = 0.25 and p = 0.97 with 500 random permutations), as evidenced by the scattered diagram (Fig. 6).

Genetic variability and heterozygosity

Table 4 illustrates the genetic variability at thirty four loci within the thirteen populations studied. The number of heterozygotes in each population was always lower than the expected number from Hardy-Weinberg equilibrium. The differences are not significant in Fito, Lanz and Pontevedra while significant in the other ten populations studied (T test, P < 0.05). The highest level of heterozygosity (H = 0.123; P = 38.2) is found in Oviedo, while the lowest (H = 0.053; P = 26.5) in Montseny. Effective size of the populations highly influences heterozygosity values, but such data were not available for our populations.

F-Statistics (Wright, 1965; 1978)

The Fst average is 0.403 for the thirteen populations. This value indicates that less than half of the measured genetic variance of these populations is partitioned among populations. The Fis parameter can be viewed as an estimate of the level of inbreeding in the populations. In our data set, the Fis parameter showed an average positive value of 0.388 (from 24 polymorphic loci, just 4 showed negative values, i.e. an excess of heterozygote while all the others exhibited positive values indicating a deficit in heterozygotes).

Discussion

The results of this study strongly support the hypothesis that the viviparous and ovoviviparous populations of Sulumandrru salumandru are conspecific. The genetic distances among the populations sampled (D,,; from 0.05 to 0.199), fall well within the intraspecific ranges reported for other salamandrids and urodeles in general (for example, see Nascetti et al., 1988; Good and Wake, 1992; Highton et al., 1989; Highton, 1990; Macgregor et al., 1990; Tilley et al., 1990; Arano et al., 1991). The absence of fixed alleles in our populations suggests the presence of gene flow and constitutes further support for the single species hypothesis. In some plethodontid species, at the intraspecific level, a high degree of life history variation with a low level of gene flow is sufficient to maintain the conspecificity of the populations (Tilley and Bernardo, 1993).


Traditionally, the populations of Sulumundra salamandra in the Northern Iberian Peninsula have been taxonomically subdivided into four different subspecies based on coloration patterns (Mertens and Muller, 1940; Thorn, 1968) blood protein analysis (Gasser, 1975; Fachbach, 1976; Gasser, 1978a), and in the case of S.S. hernardezi, the strict occurrence of direct development (viviparity) (Wolterstorff, 1928; Hillenius, 1968; Joly, 1968; Thiesmeyer and Haker, 1990; Dopazo and Alberch, 1994) (Fig. 2). We can update this scheme by proposing two well corroborated taxonomic groupings: A ~ “terrestris” and “gallaica”; B - “bernardezi” and “fastuosa”. This division correlates with phenotypic characteris-

tics (striped (B) versus blotched (A)) and reproductive modes (ovoviviparity (A) versus viviparity, including populations where this mode is not fixed (B)).

A systematic revision is beyond the scope of this paper (comprehensive studies on the biogeography and systematics of Iberian Salamandra based on allozyme data (Alcobendas, Garcia-Paris and Alberch, submitted) and DNA sequence variability (Garcia-Paris, Alcobendas and Alberch, submitted) have been recently completed). The pattern of relationship outlined in this paper, however, allows us to advance two hypothesis: one, the four subspecies involved in our study (terrestris, gulluica, hernardezi and,fastuosu) most likely are paraphyletic assemblages without biological meaning as evolutionary entities; and two, the data are consistent with viviparity evolving just once in the group.

The clear lack of correlation between genetic and geographic distances points towards a complex evolutionary history for the group in the region (Fig. 6). That is, we are not dealing with a straightforward process of gradual radiation and diffusion throughout the region as evolutionary diversification proceeded. The resulting pattern of populational variability and lineage relationships results from a combination of biological characteristics that affect the dynamics of population genetics parameters, as well as the presence of geographical barriers (Pyrenees and Cantabrian mountain ranges). Furthermore, the process is complicated by a complex sequence of secondary contacts that resulted from two distinct evolutionary “radiations”, each being associated with asynchronous colonization events, correlated with the history of glacial activity in the region (Lopez-Martinez, 1989; Real-Gimenez, 199 1).

The first ingredient in our reconstruction of the evolutionary history of S. salamandra in the Northern Iberian Peninsula involves the species specific populational dynamics as indicated by heterozygosity levels and Fst values. Low levels of gene flow and high values of Fst result in the regional fragmentation of the species into units composed of a single population, or groups of populations that remain isolated during long periods of time (Larson et al., 1984). This feature effectively allows the isolated units to evolve independently, as argued by Gill ( 1979) for the salamandrid Notophthalmus viridescens and Larson ( 1984) for various species of plethodontid urodeles. The average Fst value (0.403) of S. salamandra in the Iberian Peninsula is within the range of values observed for some Plethodontidae salamanders (Larson, 1984; Wake and Yanev, 1986; Good et al., 1987). For Triturus itulicus, the value of Fst, for 11 studied populations, was 0.448 (Ragghianti and Wake, 1986). Following these authors, we assume that the simultaneity of low

Allozyme variation in S. .w/rrmundrrcr 99

genetic distance and high Fst values indicates that part of the species has been isolated with little or no gene flow for extensive periods of time. This situation is also observed in the fragmentation of Sulumundru s&mun~lru described above.

Nevo ( 1978) reports heterozygosity levels ranging from 0.033 to 0.109 for some species of urodeles. For Rhyucotriton (Good et al., 1987) and Ensutinu eschscholtzii (Wake and Yanev, 1986) this parameter ranges from 0 to 0.145 and from 0.019 to 0.250 respectively. Therefore, the heterozygosity found in S. xdumundru populations (from 0.055 to 0.123) does not diverge from the range of values observed in other species of urodeles. Moreover, a tendancy towards heterozygote deficit is significant for 10 of the 13 populations sampled (Table 4). When the Hardy-Wein- berg equilibrium is modified, and associated with an heterozygote deficit, the mating system of the species is characterized by inbreeding (Crow and Kimura, 1970). Such a modified equilibrium observed in 10 of our 13 populations suggests that in Sulumundru sulumundru, the heterozygote deficit had to be the result of an increase in inbreeding. Furthermore, the always positive Fis values suggest non random mating.

As mentioned previously, besides populational traits such as inbreeding and limited dispersal, the gene pool fragmentation is accentuated by historical factors and geographic barriers. The existence of two separate radiations in the region was already suggested by Gasser ( 1987b), who argued that they occurred as a conse- quence of the Iberian Peninsula acting as a refugium during glacial periods. Our data do not enable us to comment on the specific timing of the two events. We propose that the viviparous populations from Asturias (“hernardezi”) and their Eastern neighbors from Cantabria, Navarra and Southwestern France exhibiting intrapopulational variability in reproductive modes (“j&tuosu”) represent a separate radiation. The other lineage -reflection of an independent historical event- includes strictly ovoviviparous forms that have been ascribed to the subspecies “ttwestris” and “gulluica”.

Such a scenario clarifies some apparently paradoxical results. Thus, when geographic distances are superimposed on our hypothesis of lineage relationships derived from the Continuous Maximum Likelihood method, we observe that the highest genetic distance occurs between the populations of Arguenos and Lindus, (D, = 0.199) which are separated by a relatively short geographic distance, but one that includes the Pyrenean mountains (Fig. 5a). Conversely, the populations of Montseny and Arguenos, also separated by the Pyrenean mountains, and a longer geographic distance, show a D, of just 0.034. On this basis, we may suggest that the Northern and Central European Sulumandru are closer relatives to our group A. Thus, we tentatively postulate a wider radiation of group A throughout the Northern Peninsula (e.g. Montseny/Pontevedra separated by 887 Km, and just a D, of 0.025) and spreading into France. The “viviparous” group is isolated into a relatively small enclave bounded by the Atlantic ocean and the Cantabrian and Pyrenean mountains. This geographic isolation favoured an independent evolution of these groups.

In conclusion, the viviparous populations of Sulumundru sulumundru from North- ern Iberia are part of a distinct lineage that includes populations exhibiting strict


viviparity as well as others showing variability in reproductive modes. The viviparous lineage represents an independent, and probably earlier radiation, of the species in Northern Spain. The data shown in this paper suggest the existence of exchange between the two basic gene pools. A more detailed analysis of the dynamics of genetic interchange in the areas of secondary contact is postponed to the second part of this study where we integrate the data reported here with new results on geographic variability of mtDNA haplotypes (Dopazo, Boto and Al- berth, submitted). The insights provided by the survey of microevolutionary patterns of geographic variation in molecular traits introduced here and in the accompanying papers provide an evolutionary background against which we can address the processes of evolution of life history strategies, and the genesis of terrestrial reproduction (“viviparity”), in this species of urodeles.

Acknowledgements

The authors wish to express their gratitude to E. Reoyo and E. Aporta for their efficient technical assistance. M. J. Blanco, M. Garcia Paris and J. Cifuentes aided us in collecting specimens. The two anonymous referees and Prof. G. Hewitt (editor of the paper) improved the manuscript with their comments and suggestions. This research was supported by the General Directorate for Science and Technology of Spain (DGICYT) grants PB X9-0045 and PB 91-0091 to P. Alberch (P.I.), by a

post-doctoral fellowship (Ministry of Education and Science of Spain) to M. Alcobendas and by a pre-doctoral fellowship from the cooperation program with Ibcro-America (Ministry of Education and Science of Spain) to Hernan Dopazo.

References

Arano, B., J. W. Arntzen, P. Herrero and M. Garcia-Paris. 1991. Genetic differentiation among Iberian

populations of the Alpine newt, 7kifurus alpestris. Amph. Rept. 12. Bas, S. 1983. Estudio de la situacibn microevolutiva y de la ecologia dc Sul~zmonu’ru suluman& (L.) en

el N.OE. ibcrico. Thesis. (Univ. Santiago, Spain) 295 pp. Blackburn, D. G. 1994. Discrepant usage of the term “ovoviviparity” in the herpetological literature.

Herpetol. J. 4: 65572. Crow, J. F. and M. Kimura. 1970. An introduction to population genetics theory. Harper & Row (New

York) 591 pp. Dopazo, H. and P. Alberch. 1994. Preliminary results on optional viviparity and intrauterine siblicide in

S&tnnn& salumundra populations from Northern Spain. Mertensiella 4: I25 137.

Duellman, W. E. and L. Trueb. 1986. Biology of the amphibians. McGraw Hill Book Comp. 670 pp. Eiselt, J. 1958. Der Feuersalamander, Salumandra .salumandru (L.). Beitrage zu einer taxonomischen

Synthesc. Abh. Ber. Mus. Naturak. u. Vorgesch. (Magdeburg) IO: 77 154. Fachbach, G. 1971. Zur Klarung verwandtschaftlicher Beziehungen bei Vertretern der Gattung Salu-

mundru mit Hilfe der Polyacrylamid-Disk-Electrophorese. II. Z.f. Zoo]. Syst. u. Evolut. Forschung 9: 181~ 187.

Fachbach, G. 1976. Biologie, Taxonomie und phylogenetiache Beziehungen dcr verschiedenen Unter-

arten von Sulumandru .saiamundra im Bereich der Ibcrischcn Halbinscl. Z.f. Zool. Syst. u. Evolut. Forschung 9: 49 60.

Farris, J. S. 1972. Estimating phylogenetic trees from distance matrices. The Amer. Natural. 106: 645 -668.

Felsenstein, J. 1989. PHYLIP-Phylogeny infcrcnce package (version 3.2). Cladistics 5: 164 166.

AlloTyme variation in S. .vultrmtm&u 101

Garcia-Paris, M. (1985). Los anfibios de Espafia. Madrid, Publicaciones de extension agraria

Gasscr, F. 1975. Recherchcs sur la statut microevolutif de deux amphibiens Urodcles. I’espece pyre- neenne EuP~~~/u.F rasper (Dug&s) et l’csptce palearctique Strlamrm&l .vtrklmrrntlrcr (L.): proteincs et groupes seriques, cycles sexuels femellcs et morphologie. Thesis AO. CNRS 121 12. (Univ. Paul

Sabatier, Toulouse, France) 267 pp. Gasser, F. 1978a. Le polytypisme de l’espece paltarctique S&ntrmtlrtr .str/umrm~lrrr (L.) ( Amphibicn,

Urodele). I. Proteines seriqucs et groupes seriques. Arch. Zool. exp. g&n. 119: 585 617. Gasser, F. 1987b. Le polytypisme de l’espece paltarctique Salrmrcmu’rr~ .scr/cr,rrtrrr&cl ( L.) ( Amphibicn

Urodele). II. Systematique. relations g&i-tiques et tendances cvolutives dans l’airc de repartition. Arch. Zool. exp. g&n. 119: 635 66X.

Gill, D. E. 1979. Density dependence and homing behavior in adult red-spotted newts Notophthtrln?u.r viri&scens (Rafinesque). Ecology, 60: 800 8 13.

Good, D. A. and D. B. Wake. 1992. Geographic variation and speciation in the torrent salamanders of

the genus Rhycccotriton (Cuudurrl: Rhq’“‘OtrifoiziLZuc). Univ. California Publ. Zoo]., 126: 1 9 1. Good, D. A., G. Z. Wurst and D. B. Wake. 19X7. Patterns of geographic variation in allozymca of the

Olympic Salamander Rhyzcotriton olympicu.~ (Cauiufu: ni~c/llll,rotlonf/rklc,). Fieldiana, Zool. 32:

I-15. Greven, H. and B. Thicsmeier. 1994. Biology of Sulrrtntm~lru and Mertens~elkr. (Proc. Sympos. Biology

of Salumcmdrcr and Mcrtcn.sie/la, Dusseldorf). In Mertensiella Suppl. 4, 454 pp. Harris, H. and D. A. Hopkinson. 1976. Handbook of Enzyme Electrophorcsis in human genettcs.

North-Holland Pub]. Co., (Amsterdam). Highton, R. 1990. Taxonomic treatment of genetically differentiated populattons. JIerpetologica. 46( I):

114 121. Highton, R., G. C. Maha and L. R. Maxson. 1989. Biochemical evolution in the slimy salamanders of

the P&ho&n &&no.vus complex in the Eastern United States. Univ. Illinois Biol. Monogr. S7: IS3

PP. Hillenius, D. 1968. Notes on Sulumcmdrrr scdumtmdru ssp. Bydr. Dierk., 3X: 31 -38. Hillis, D. M. 1985. Evolutionary genetics of the andean lir.ard genus Phol/rk&olus (.%urI’u: C;l’!?rno(,lr-

thrdmidw): Phylogeny, biogeography, and a comparison of tree construction techniques. Syst. Zool. 34: 109m 126.

Joger, U. S. y Steinfar/. ( 1994): Zur subspezifischen Glicdcrung der stidiberisschcn Feuersalamander. Abh. bet-. Naturkde. Vorgesch. Magdeburg 17: 83398.

Joly, J. 1968. Don&es tcologiques sur la salamandre tachette Srr/mlnn&r .sc~/crmcm& (L.). Ann. Sci. Nat. Zoo]. IO: 301 -366.

Klewen, R. 1991. Die Landsalamandcr Europas, Teil 1 Die ncue Brehm-Bticheret 584, 2nd ed.. Wittenberg-Lutherstadt.

Larson, A. 1984. Neontological infcrcnces of evolutionary pattern and process in the salamander ?dmily

Plethodontidw. Evol. Biol. 17: 119~ 217. Larson, A., D. B. Wake and K. P. Yanev. 1984. Measuring gene how among populations having htgh

levels of genetic fragmentation. Genetics 106: 293-308. Lopez-Martinez, N. 1989. Tendencias en Paleobiogeografia. El futuro de la biogeogralia. Paleontologia.

Coll. Nuevas Tendencias CSIC (Madrid, Spain), 271 296. Macgregor, H. C., S. K. Session and J. W. Arntzen. 1990. An integrative analysis of the phylogcnetic

relationships among newts of the genus Triturus (family Salamandridae), using comparative

biochemistry, cytogenetics and reproductive interactions. J. evol. Biol. 3: 329 373. Malkmus, R. 1983. Beschreibung ciner neuen Form des Fcucrsalamanders der Serra de Monchique

(Portugal): Suirmun&~ scdumundro (&/aica) crespoi n. subsp. Faun. Abh. Mus. Ticrk. Dresden 10(9): 1699174.

Mertens. R. and L. Miiller. 1940. Die Amphibien und Reptilien Europas (2 listc. nach dem stand vom I. Januar 1940). Abh. senckcnberg. naturf. Ges.. Frankfurt am Main, 451: I-56.

Nascetti, G., F. Andreonc. M. Capula and L. Bullini. 1988. A new S~dumcmdrrr species from southwcst- ern alps (Amphihra, (/rod&, S~dumundrrido~). Boll. Mus. reg. SCI. Nat. Torino. 6: 617-638.


Nei, M. 1977. F-statistics and analysis of gene diversity in subdivided populations. Ann. Human Genet. 41: 225 233.

Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590.

Nevo, E. 1978. Genetic variation in natural populations: patterns and theory. Theor. Popul. Biol. 18: I21 ~177.

Nitecki, M. H. 1990. Evolutionary innovations. The Univ. Chicago Press (Chicago): 304 pp. ijzeti, N. 1979. Reproductive biology of the salamander Mer/ensiel/u luschuni utrtc~/yunu. Herpetologica

35(3): l93- 197. Pasteur, N., G. Pasteur, F. Bonhomme. J. Catalan, and J. Britton-Davidian. 1987. Manuel de genctique

par tlectrophorese des proteines. Coll. Technique et Documentation, (Lavoisier, Paris) 217 pp. Ragghianti, M. and D. B. Wake. 1986. Genetic variation and its evolutionary implications in the Italian

newt, Triturus Itcdicus. Herpetologica 42: 206-214. Real-Gimenez, R. 1991. Modelos de distribution de 10s anfibios en las cuencas fluviales de Europa a trcs

escalas geogrificas. Thesis. (Univ. Malaga, Spain) 374 pp. Rohlf, F. J. 1993. NTSYS-pc Numerical Taxonomy and Multivariate analysis System, Version 1.80.

Setauket. New York. Schrciber, E. 1912. Ilerpetohgia eurc)/~~ra. 2nd ed. G. Fischer. Jena.

Sclander, R. K., M. H. Smith. S. Y. Yang, W. E. Johnson and J. B. Gentry. 1971. Biochemical polymorphism and systematics in the genus Peromyscu.s. I. Variation in the old field mouse (Prromy.vcu.r pohnotus). Studies in Genetics, (Univ. Texas Pub].) 7103: 49-90.

Shaffcr, H. B., J. M. Clark and F. Kraus. 1991. When molecules and morphology clash: A phylogenetic analysis of the north american ambystomatid salamanders (Cuudutu: Amhystomatidae). Syst. Zool. 40( 3): 2844303.

Swofford, D. L. 1981, On the utility of the distance Wagner procedure. In V. A. Funk and D. R. Brook (eds). Advance in cladistic, Proceed. of first meeting of the Willi Hennig Sot. (New York).

Swofford. D. L. and R. B. Selander. 1981. Biosys-I, a computer program for the analysis of allelic variation in genetics. (Univ. Illinois, Urbana).

Thiesmeier, B. and K. Haker. 1990. Sulrrmcmdru scrlumandru hrmurde-i (Wolterslorff, 1928) aus Oviedo. Spanien. ncbst Bemerkungen zur Viviparie in der Gattung Solrmundru. Salamandra 26: 140- 154.

Tilley. S. G. and J. Bernardo. 1993. Life history evolution in plethodontid salamanders. Herpetologica

49(2): 1544163. Tillcy. S. G., P. A. Verrell and S. J. Arnold. 1990. Correspondence between sexual isolation and

allozyme differentiation: A test in the salamander, I)r.snzo~nulhus ochrophuws. Proc. Nat. Acad. Sci. (USA) 87: 2715 -2719.

Thorn, R. 1968. Lcs salamandres d’europc, d’asie et d’afrique du nerd. Paul Lechevalier 35 (Paris). Wake. D. B. and K. P. Yanev. 1986. Geographic variation in allozymcs in a “ring species”, the

plethodonttd salamander Ensatinu cwhscholtzii of western north america. Evolution 40: 702 715. Wake. M. 11. 1982. Diversity within a framework of constraints: reproductive modes in amphibians. In

D. Mossakowski and G. Roth (eds.), Environmental adaptation and evolution, a theoretical and empirical approach, 87 106 (G. Fischer).

Wake. M. H. 1993. Evolution of oviductal gestation in amphibians. J. Exp. Zool. 266: 3944413. Warburg, M. R., G. Degani and I. Warburg. 1979. Ovoviviparity in Sulrrmundru .s&mundra (Arnphibiu,

Urod~lc~) from Northern Israel. Vie Milieu 28/29: 2477257. Wolterstorff, W. 1928. Vollmolch-geblrende Feuer-salamander aus Oviedo. Bl. Aquar. Terrar. Kde. 39:

1322133. Wright. S. 1965. The interpretation of population structure by F-statistics with special regard to systems

of mating. Evolution 19: 395 420. Wright, S. 1978. Evolution and the genetics of populations. Vol. 4: Variability within and among natural

populations. (Univ. Chicago Press, Chicago, 11).

Received 19 October 1994; accepted 13 July 1995. Corresponding Editor: G. M. Hewitt

geographic variation in allozymes of populations of salamandra

Documents