Sunteți pe pagina 1din 8

J. Zool., Lond. (2002) 257, 155±162 # 2002 The Zoological Society of London Printed in the United Kingdom DOI:10.


Phenotypic plasticity of anuran larvae: environmental variables

in¯uence body shape and oral morphology in Rana temporaria

Miguel Vences1*, Marta Puente2, Sandra Nieto2 and David R. Vieites2

MuseÂum national d'Histoire naturelle, Laboratoire des Reptiles et Amphibiens, 25 rue Cuvier, 75005 Paris, France
Laboratorio de AnatomãÂa Animal, Facultade de Ciencias BioloÂxicas, Universidade de Vigo, Apdo. 874, 36200 Vigo, Spain
(Accepted 23 May 2001)

Environmental variables shaped the morphology of tadpoles of the common frog, Rana temporaria, in
various ways at the Pyrenean locality Circo de Piedra®ta. Examining only specimens in similar
developmental stages, those from small ponds (with higher temperature and higher tadpole density) had
lower growth rates, lower relative tail height, lower relative body width and fewer labial keratodonts and
keratodont rows. The variation in keratodonts may have been caused by heterochrony related to the
slower growth rate. The number of lingual papillae also differed between ponds but was not related to
pond size. Higher predator densities caused a higher percentage of damaged tails and a lower relative tail
length in specimens with apparently intact tails, probably as a result of incomplete regeneration after
mutilations earlier in development.

Key words: Amphibia, Ranidae, Rana temporaria, tadpoles, phenotypic plasticity, pond size, predator
density, oral structures

INTRODUCTION known to occur in Hyla chrysoscelis tadpoles

(McCollum & Leimberger, 1997). A number of morpho-
Time to metamorphosis and rate of development for logical variables and colour pattern differences between
anuran tadpoles is in¯uenced by the temperature brook-dwelling and pond-dwelling larvae of Rana chir-
(Harkey & Semlitsch, 1988), size (Pearman, 1993) and icahuensis are known to exist (Jennings & Scott, 1993),
rate of dessication (Newman, 1989) of their habitat while the number of oral papillae may differ depending
(Denver, 1997), plus density (e.g., Gromko et al., 1973, on environmental variables in Rana berlandieri and
Miranda & PisanoÂ, 1993), pathogens (Beebee, 1995; R. sphenocephala (Hillis, 1982). Although a low percen-
Petranka, 1995) and diet (Kupferberg, 1997). Differ- tage of specimens with aberrant oral morphology are
ences in developmental rate are also caused by genetic generally present in tadpole populations (e.g. Bresler &
factors in at least some species (Berven, 1987; Blouin, Bragg, 1954), a high proportion of intra- or interpopula-
1992). There is less information published on the mor- tional variation of such features is mostly considered as
phological plasticity of amphibian larvae with the an exception. Field guides usually use tadpole mor-
possible exception of size differences. Crump (1989) and phology as species-speci®c, i.e. genetically determined.
Inger (1992) reported morphological differences seen in Vences, Kupfer et al. (1998) noted important inter-
conspeci®c feeding and non-feeding tadpoles, while geo- populational variation in the morphology of tadpoles of
graphic variation in mouthpart morphology of the common frog Rana temporaria. The systematics of
Scaphiopus tadpoles is known from North America Pyrenean brown frogs were not well assessed until
(Potthoff & Lynch, 1986). In salamanders (Pfennig & recently (see Serra-Cobo, 1993; Vences, Kupfer et al.,
Collins, 1993; Nishihara, 1996) and anurans (Pfennig, 1998), and differences in the number of labial kerato-
1992), cannibalistic morphs, which differ drastically dont rows and lingual papillae have been considered as
from non-cannibalistic morphs in head morphology, indications of taxonomic distinctness of the respective
can occur under certain environmental conditions. populations (Vences, Kupfer et al., 1998). A second
Predator-induced morphological changes are also possible explanation given by the same authors was a
possible heterochronic reduction of tooth rows caused
*All correspondence to: M. Vences. Zoologisches Forschungsinstitut by increased developmental rates.
und Museum A. Koenig, Adenauerallee 160, 53113 Bonn, Germany. During recent ®eldwork in the Spanish central Pyr-
E-mail: enees, large morphological variations of body shape and
156 M. Vences et al.

oral morphology were observed among tadpoles inha- sulphate, < 200 mg/l. Carbonate hardness and total
biting neighbouring ponds. To analyse the interaction hardness were  18 dH in all ponds except I (28 dH), J
of different ecological in¯uences producing this varia- (38 dH) and F (48 dH). Correlations of conductivity with
tion, assessments were made of the biotic and abiotic total and carbonate hardness and with pH were signi®-
characteristics of 14 ponds and the morphology of the cant (rs = 0.7670.88, P < 0.005). Such values were to be
tadpoles inhabiting them. expected in the largely granite Circo de Piedra®ta. The
presence of high amounts of other pollutants (e.g.
organic compounds) in some ponds while absent in
MATERIALS AND METHODS others, is not probable in this remote and largely
unaltered area. Therefore, it can be excluded with
Abiotic pond characteristics suf®cient probability that the observed differences in
tadpole growth rate and shape were caused by pollution
Fieldwork was carried out in the Circo de Piedra®ta (Rowe, Kinney & Congdon, 1998).
area, central Pyrenees, AragoÂn, Spain, in July 1999.
This plain of 2±3 km2 at c. 2100±2250 m elevation
includes a number of large glacial ponds, smaller ponds Biotic pond characteristics
and stagnant brook tributaries, which are used for
reproduction by a large Rana temporaria metapopula- Aquatic vegetation (Sparganium angustifolium) was
tion. The 14 ponds surveyed, labelled A±N, were present in ponds A, C, and D and was dense in pond
located between 42848.35'N±42849.04'N and 0815.59'W± G. Ponds F and J were vegetated at 30% and 80% of
0817.77'W. their surfaces, respectively, by Carex fusca. All other
Ponds were classi®ed (Table 1) as large (A±D) with a ponds had no aquatic or semi-aquatic vegetation.
water depth > 50 cm, and small (E±N). Several small Various methods that were dependent on the size and
ponds were not permanent; pond N dried out at the accessibility of the pond were used to estimate the
beginning of July, killing all tadpoles (only a limited number of tadpoles present. Direct counts were pre-
number of morphological variables could therefore be ferred where possible, but in large ponds and those with
measured in the specimens from this pond). In 1998, the many hiding places, the number was estimated by
year before we carried out the systematic ®eldwork, neutral red staining (c. 5 min in 0.025±0.1% neutral red
pond H also dried out before the tadpoles metamor- solution; see Cooke, 1976) of a large sample of tadpoles
phosed. At the end of the study period, ponds F, J per pond and recapturing tadpoles the next day
and L were also beginning to dry up. Although most (Petersen, 1896). The dye of the tadpoles was always
ponds were stagnant, pond M was connected to a small recognizable 1±2 days after staining and in some ponds
permanent brook, and pond K was a slow-running part even after 5±6 days. Tadpole mortality during the
of that same brook. All ponds were totally sun-exposed. staining procedure was low and mainly caused by a lack
The bottom of the ponds were stone or gravel, often of oxygen when too many tadpoles were stained at once.
largely covered by mud. Measurements were taken of Tadpole densities/m3 were correlated with water tem-
minimum and maximum pond diameter, circumference perature (Spearman's rank correlation, rs = 0.78,
and maximum depth, and an estimate of pond surface P < 0.005) and inversely with water volume (rs = 70.59,
was calculated. As most ponds had a more or less P < 0.05).
regular depth over most of their surface, water volume During nocturnal searches along the pond shores 5
was roughly estimated using pond surface and different potential tadpole predators were observed: (1)
maximum depth as parameters. newts Euproctus asper; (2) salamander larvae Sala-
Diurnal water temperature (measured in all ponds mandra salamandra; (3) dragon¯y larvae; (4) Dytiscus
between 16:00 and 17:00 h on 16 July 1999 to exclude beetles; (5) Dytiscus beetle larvae. No ®shes were seen in
diel or temporal variations) was highest in the shallow the ponds, although 2 species occured in the nearby
ponds (23.5±30.5 8C), lowest in the larger ponds dams (Salmo trutta, Phoxinus phoxinus). Aquatic shrews
(18.5±20.4 8C) and intermediate in K and M, which Neomys fodiens were observed sporadically in pond A
partially had ¯owing water (22.3±24.1 8C). Temperature but apparently did not occur in high densities. Direct
was inversely correlated with water volume (Spearman's observations were made of predation attempts by sala-
rank correlation, rs = 70.58, P < 0.05) and depth mander larvae on tadpoles in pond H. The salamander
(rs = 70.72, P < 0.005). larvae were not able to catch the tadpoles, which were
Water samples were preserved with a small amount of as large or even larger than the larvae, but did succeed
chloroform, and chemical parameters were studied with to bite them from behind on several occassions. The
Sera#, Aquamerck#, and Merckoquant# test kits. effect of these bites could not be ascertained in detail,
Water in all ponds was low in ions as re¯ected by low but the larvae were probably able to remove parts of the
conductivity values (Table 1) and was slightly acidic tadpole's tails. Predator density was estimated from our
(pH values 5.0±7.0). Other chemical parameters were nocturnal counts as specimens/m of shoreline; it ranged
close to or below the sensitivity limit of the tests: from 0 to 3.09/m. Numbers of salamander larvae were
calcium,  10 mg/l; nitrate, < 5 mg/l; chloride, highest in pond H (170 specimens along the shoreline).
< 12.5 mg/l; phosphate, < 0.25 mg/l; iron, < 0.05 mg/l; There were no signi®cant correlations with predator
Table 1. Main characteristics of ponds studied and selected morphological traits of the respective tadpole populations. T, water temperature; C, water conductivity; SVL, snout±
vent length of tadpoles; SVL met., snout±vent length of recently metamorphosed specimens; NPAP, number of oral papillae; K1, K2, K3, number of keratodonts of ®rst, second
and third interrupted upper keratodont rows (right side of jaw sheaths only). Measurements, mean ‹ standard deviation (minimum and maximum values in parentheses);
developmental stages, according to Gosner (1960), mean with minimum and maximum values in parentheses; tadpole density, m3 of water; predator density, potential predator
specimens/m of pond shoreline; last column, percentage of individuals with clearly damaged tails (tail tip missing or severely mutilated)

Tadpole Predator
Surface Volume density density T SVL met SVL Gosner Mutilated
Pond (m2) (m3) (per m3) (per m) (8C) C (mS) (mm) (mm) stage NPAP K1 K2 K3 tails (%)

A 9500 19000 12 0.09 19.0 40 14.54 15.4 ‹ 01.14 37 85.8 ‹ 10.1 55.0 ‹ 7.8 29.6 ‹ 11.0 11.3 ‹ 3.5 0
(12.8±17.2) (36±38) (69±107) (40±65) (0±45) (7±20)
B 1650 3150 2 0.04 18.5 12 14.13 16.35 ‹ 0.98 37 101.4 ‹ 12.8 55.4 ‹ 9.72 30.0 ‹ 9.0 8.9 ‹ 5.2 27
(14.3±17.8) (34±40) (85±127) (42±65) (14±45) (0±15)
C 1065 735 11 0.33 20.4 19 13.8 14.16 ‹ 1.95 33 91.1 ‹ 18.7 45.2 ‹ 7.3 28.1 ‹ 6.1 3.1 ‹ 3.8 42
(10.3±17.2) (28±36) (69±134) (36±60) (16±38) (0±10)
D 660 858 1 0.17 20.0 12 ± 13.89 ‹ 1.13 33 88.7 ‹ 14.2 44.9 ‹ 6.1 25.0 ‹ 3.9 7.6 ‹ 4.3 ) 53

Phenotypic plasticity of Rana tadpoles

(11.9±15.6) (30±35) (65±112) (32±54) (19±30) (0±15
E 135 34 26 0.49 23.5 12 13.22 14.03 ‹ 0.85 37 105.9 ‹ 24.8 46.3 ‹ 6.2 25.1 ‹ 4.8 2.8 ‹ 3.8 31
(12.9±15.1) (35±38) (69±160) (36±55) (18±35) (0±10)
F 180 49 30 0.09 25.6 100 12.9 12.13 ‹ 1.31 34 76.6 ‹ 21.1 39.8 ‹ 7.9 14.6 ‹ 6.7 0 13
(9.6±13.5) (25±41) (26±98) (26±50) (5±25)
G 75 8 181 0.24 29.1 32 13.5 14.19 ‹ 1.09 34 100.6 ‹ 13.8 45.0 ‹ 6.9 24.0 ‹ 5.9 0.2 ‹ 0.6 25
(12.9±16.5) (31±37) (83±126) (35±55) (15±30) (0±2)
H 215 43 211 3.09 27.1 18 12.23 13.31 ‹ 1.04 35 94.3 ‹ 18.9 46.9 ‹ 10.5 24.6 ‹ 6.9 0.5 ‹ 1.3 47
(11.2±14.6) (26±41) (63±116) (32±65) (14±35) (0±4)
I 30 9 632 0 26.4 62 10.12 9.81 ‹ 1.68 30 90.4 ‹ 19.3 44.8 ‹ 6.34 13.1 ‹ 6.6 0 0
(7.5±13.0) (25±41) (65±116) (35±50) (3±20)
J 70 8 4 0 25.0 79 ± 12.18 ‹ 0.64 37 89.7 ‹ 21.7 43.5 ‹ 5.2 22.8 ‹ 6.9 1.2 ‹ 2.1 17
(11.1±13.1) (36±41) (66±130) (40±57) (15±35) (0±5)
K 10 1 163 0 24.1 ± ± 11.09 ‹ 1.92 29 76.0 ‹ 7.9 35.9 ‹ 3.9 19.2 ‹ 2.3 0.3 ‹ 1.0 ) 14
(9.9±17.5) (25±32) (65±93) (30±43) (16±23) (0±3
L 5 1 41 0 25.1 30 11.2 11.45 ‹ 1.31 34 106.7 ‹ 11.1 39.6 ‹ 3.6 25.8 ‹ 5.0 5.8 ‹ 3.9 1
(8.4±13.2) (26±36) (76±118) (33±45) (16±35) (0±10)
M 2 1 30 0 22.3 25 ± 12.08 ‹ 1.48 32 86.8 ‹ 12.8 36.9 ‹ 5.5 21.1 ‹ 5.5 0 0
(10.7±14.9) (29±35) (70±111) (31±50) (16±30)
N 20 1 ± 0.27 30.5 ± ± 10.59 ‹ 0.82 37 ± 33.3 ‹ 14.2 14.8 ‹ 10.5 0 ±
(9.5±12.0) (37±38) (15±48) (0±24)

158 M. Vences et al.

density of any physical or chemical variables, nor with

tadpole density.
No data are available on the start of the breeding
season of frogs in the different ponds in the study year.
However, data from May and June 2000 showed that
breeding started more or less simultaneously in different
small and large ponds of the Circo de Piedra®ta. Clutches
were already present in some of the smaller ponds (e.g.
pond I) heated by the sun, while they were laid some
days later in the large pond A. Thaw most probably
proceeds each year in a similar succession in the ponds
of the area. It can therefore be excluded with suf®cient
probability that tadpoles from small ponds were consis-
tently younger than those from the large ponds.

Test of genetic homogeneity

To verify that the morphologically different tadpoles (b)

belong to the same species (see Vences, Kupfer et al.,
1998), a fragment (up to 544 bp) of the mitochondrial
16S rRNA gene was sequenced in 1 specimen from
ponds A±C, E±F, H±I and L±M. Primers, PCR proto-
cols and sequencing conditions were as in Vences,
Kosuch et al. (2000). No differences were detected
among the sequences (Genbank accession numbers
AF275735±AF275743) except a single substitution in
the tadpole of pond C at position 208 of the alignment.
Western Palearctic brown frog species differ by least 3%
of the homologous gene fragment (M. Veith, J. Kosuch
& M. Vences, pers. obs.). The genetic data thus con-
®rmed that the different tadpole morphs correspond to
a single species, Rana temporaria.

Fig. 1. Different types of oral morphology observed in Rana

Morphological measurements temporaria tadpoles of the Circo de Piedra®ta. Note: three
interrupted upper tooth rows in specimen from pond A (a),
Tadpoles (10±15 specimens) were collected in each pond two interrupted upper tooth rows in specimen from pond I (b).
on 15 and 16 July 1999, anaesthetized and preserved in
4% formalin. Terminology of tadpole morphology
follows Altig & McDiarmid (1999) and Viertel & pointed) tail tip. One to 3 specimens of each sample
Richter (1999) except for the term `keratodont', which were prepared by critical point drying and their mouth-
we use for the labial teeth of tadpoles according to parts, especially buccal ¯oor and buccal roof, were
Dubois (1995). The developmental stage was recorded examined with a scanning electron microscope (SEM).
according to Gosner (1960); snout±vent length (SVL), All specimens (except those used for SEM analysis) are
tail length (TL), maximum tail height (TH), maximum preserved in the collection of the Zoologisches For-
body width (BW), and number and arrangement of schungsinstitut und Museum A. Koenig, Bonn,
keratodont rows using the standardized formula of Germany (ZFMK 72115±72140). Statistical analysis
Dubois (1995), number of buccal papillae, and number was carried out with SPSS for Windows (version 9).
of keratodonts in the ®rst, second and third interrupted
upper rows (K1, K2, K3) on the right side of the jaw
sheaths. Morphometric measurements were taken with RESULTS
callipers to the nearest 0.1 mm and mouthparts were
examined with a stereomicroscope. In some samples, a All measured morphological variables and develop-
large percentage of tadpoles had damaged tails mental stages differed signi®cantly among ponds
(compare Blair & Wassersug, 2000). This was obvious (Kruskal±Wallis ANOVA, P < 0.001). Highly signi®cant
as generally the posterior part of the tail was missing to differences (P < 0.001) were found in all variables except
varying degrees. A specimen was scored as having a for K3 (not signi®cant) and K2 (P < 0.05) by ANCOVA
damaged tail only if the mutilation was obvious from a with SVL as the covariate; and in all variables except
fresh wound or recognizable from the straight (not K1 (P < 0.005) by ANCOVA with stage as the covariate.
Phenotypic plasticity of Rana tadpoles 159

(a) (c)

Size-adjusted body width

SVL (mm) 9.5


10 7.5
1 10 100 1000 10000 1 10 100 1000 10000
(b) (d)
Size-adjusted tail height

No. of keratodonts
8.0 40


1 10 100 1000 10000 1 10 100 1000 10000
Volume (m3)
Fig. 2. Growth rate and morphological traits of Rana temporaria tadpoles relative to pond size. Snout±vent length (a), size-
adjusted tail height (b), size-adjusted body width (c), and number of keratodonts of the upper ®rst (*), second (*) and third (^)
interrupted tooth row (d) were plotted against water volumes (logarithmic scale). Each symbol represents mean value of one
pond (only specimens in developmental stages 35±38 considered). Tail height and body width were adjusted to the mean overall
SVL of 12.99 mm. Correlations were signi®cant (Spearman's rank correlation, P < 0.05) in A, C, and D, as well as for total
number of keratodonts and keratodonts on ®rst interrupted upper row in B.

Tadpole size differences between ponds were very con- and negatively correlated with tadpole density
spicuous; mean SVL in pond B was almost twice the (rs = 70.90, P < 0.005) but not with temperature.
value in pond I. Tadpole keratodont formulas differed To exclude the effects of stage differences on the
in the number of interrupted keratodont rows on the morphometric variables, mean values per pond were
upper lip. These were three (resulting in a formula of calculated based on specimens in stages 35±38 only. No
1:3 + 3/1 + 1:3) in almost all specimens of the three such specimens were present in the samples of ponds I
largest ponds (A, B, D), in about half of the specimens and K; the analyses therefore were carried out with a
in the fourth largest pond (C), and in most specimens of reduced data set of 11±12 ponds. Numbers of upper
pond L. Tadpoles in the remaining ponds had exclu- keratodont rows were similar to the overall sample. In
sively or mostly only two interrupted upper rows the four large ponds the percentage of specimens with
(formula 1:2 + 2/1 + 1:3) (Fig. 1). SEM pictures did not three upper keratodont rows was 66±80% (ponds A, B
reveal consistent differences in size and arrangement of and D) and 50% (pond C). In pond L, 92% of the
the papillae of the buccal ¯oor and buccal roof, which tadpoles had three rows while the percentages were
corresponded well to the drawings of Viertel (1982). All 0±35% in the remaining small ponds. SVL (as most
specimens had four lingual papillae; in several indivi- other morphological measurements) was correlated with
duals, one to three of these were damaged (only their water volume (rs = 0.60, P < 0.05). Mean number of
bases visible) and dif®cult to recognize. keratodonts of all three rows studied increased with
Mean Gosner stage was not signi®cantly correlated water volume, but the respective correlation was only
with water volume, temperature, nor tadpole density. signi®cant for K1 (rs = 0.74, P < 0.01). A signi®cant
Metamorphosis occurred in nine ponds during the study correlation was also found for the total number of
period. The date of ®rst metamorphosis observation keratodonts, when all three rows were added together
(developmental stages 41±46) was 14 July in pond F, 21 (rs = 0.65, P < 0.05; Fig. 2). No obvious differences in
July in ponds E and I, 22 July in pond B, 25 July in keratodont shape or size were noted in the SEM ana-
pond A, and 28 July in ponds C, G, H, and L and was lyses. Similar (but inverse) correlations and signi®cance
not correlated with water volume, temperature, or levels were found between morphology and water tem-
tadpole density. Mean SVL of the metamorphosing perature but not with tadpole density. No signi®cant
specimens (up to ®ve individuals measured) was posi- correlation of any environmental variable was noted
tively correlated with water volume (rs = 0.75, P < 0.05), with number of papillae.
160 M. Vences et al.

25 1993; Reques & Tejedo, 1995). According to Denver

(1997), the same effects may be inhibitory to growth if
Size-adjusted tail length

present early in the larval phase, but stimulatory to
development if present during prometamorphosis. Rana
21 temporaria tadpoles at the Circo de Piedra®ta probably
experienced the relevant environmental in¯uences (dif-
19 ferences in pond size, density and temperature) upon
hatching. They grew slower and metamorphosed at a
smaller size with decreasing pond volume (and in-
creasing temperature and density), although no
signi®cant modi®cation of time to metamorphosis was
15 noted. The observed pattern seems to match better the
0 0.1 0.3 1 4
Predators/m second category outlined above. Which factor or combi-
nation of factors is most responsible for the growth
Fig. 3. Size-adjusted tail length of specimens with apparently differences is dif®cult to assess without controlled ex-
intact tails plotted against predator density (individuals periments, considering their strong inter-correlations.
counted per m of pond shore). Each symbol represents mean As the ponds studied were poor in minerals and algae it
value of one pond. Values are based on specimens in stages could be assumed that food was a limiting resource in
35±38 only. Predator density, semi-logarithmic scale to ponds with high density, accounting, partially at least,
account for the high values in pond H. The correlation was for the reduced growth.
signi®cant (Spearman's rank correlation, P < 0.005). The factors that account for differences in growth
rate and size at metamorphosis among tadpoles in the
Circo de Piedra®ta ponds are also likely to have effects
on oral structures. The upper keratodont rows in Rana
To exclude the size factor from the morphometric appear centripetally (Grillitsch & Grillitsch, 1989), as in
measurements TL, TH and BW, we log-transformed other anurans (e.g. Hoplobatrachus tigerinus; Dutta &
these variables, and subsequently removed log(SVL) as Mohanty-Hejmadi, 1984). This means that the outer
a covariate through regression analysis. Values were rows appear ®rst, while the rows closest to the jaw
adjusted to the overall mean SVL (12.99 mm); thus, all sheaths sequentially form at later stages. Length of rows
specimens were compared as if they had this average and keratodont number increase during anuran devel-
SVL (Blouin, 1992). Size-adjusted tail height was corre- opment (Tubbs et al., 1993). Specimens in earlier
lated with water volume (rs = 0.648, P < 0.05) and developmental stages thus have a lower number of
inversely with tadpole density (rs = 70.651, P < 0.05). keratodonts and keratodont rows. Among the surveyed
Size-adjusted body width was positively correlated with ponds, however, such differences were also obvious in
volume (rs = 0.680, P < 0.05) and negatively with specimens of similar stages. It is assumed that hetero-
tadpole density (rs = 70.647, P < 0.05) and temperature chronic developmental shifts were responsible for these
(rs = 70.618, P < 0.05). Size-adjusted tail length of spe- observations; progress of development, and eventually
cimens with apparently non-damaged tails was not metamorphosis, being induced before the oral structures
correlated with volume, temperature and density, but are fully developed. This hypothesis most probably
was negatively correlated with predator density explains why growth-inhibited tadpoles in the small
(rs = 70.831, P < 0.005; Fig. 3). The percentage of ponds had few keratodonts in late stages.
damaged tails among complete samples of all ponds Tadpoles in the small ponds also had relatively lower
was positively correlated with predator density tail ®ns and their bodies were narrower. Almost 50% of
(rs = 0.727; P < 0.01) but not with any other environ- the dry weight of a Rana tadpole is stomach content
mental factor. (Calef, 1973). Our ®eld impressions were a general lower
body volume in specimens from small ponds as com-
pared to the almost spherical bodies of tadpoles found
DISCUSSION in large ponds. A lower food intake or weight loss
(either due to lower food availability or related to
Modi®cations of tadpole growth rate and size at meta- crowding stress or pathogen factors; Gromko et al.,
morphosis by environmental factors are of two sorts. 1973; Beebee, 1995; Petranka, 1995) in the smaller
With increasing temperature (Harkey & Semlitsch, ponds may therefore account for the variation in body
1988) or pond dessication rate (e.g. Newman, 1989; width. The differences in tail height may be related to
Denver, Mirhadi & Phillips, 1998; Loman, 1999) time to the long swimming distances covered each day by tad-
metamorphosis may be reduced and metamorphosis poles in the large ponds, which migrated each day from
occurs at smaller or similar body sizes. Alternatively, the deep part of the ponds to the shallow edges,
time to metamorphosis is prolonged, and metamor- probably due to thermoregulatory behaviour patterns
phosis nevertheless occurs at smaller sizes; this response (Pique et al., 2000).
has been observed at lower food resources or higher A further, independent modi®cation of tadpole shape
density (e.g. Gromko et al., 1973; Miranda & PisanoÂ, was caused by predation. Predators caused a higher
Phenotypic plasticity of Rana tadpoles 161

proportion of damaged tails, and a lower relative tail REFERENCES

length in specimens with apparently intact tails. Most
probably, these specimens were attacked and damaged Altig, R. & McDiarmid, R. W. (1999). Body plan: development
by predators early in development and their tails grew and morphology. In Tadpoles. The biology of anuran larvae:
24±51. McDiarmid, W. W. & Altig, R. (Eds). Chicago: The
back completely, albeit shorter. This observation would
University of Chicago Press.
be in accordance with the hypothesis that tail loss in Barbadillo, L. J., Lacomba, J. I., PeÂrez-Mellado, V., Sancho, V. &
anuran larvae incurs little cost and therefore may be an LoÂpez-Jurado, L. F. (1999). An®bios y reptiles de la PenãÂnsula
important mechanism to reduce the effect of predation IbeÂrica. Barcelona: Geo Planeta.
(Wilbur & Semlitsch, 1990). Beebee, T. J. C. (1995). Tadpole growth: is there an interference
Altogether, the morphological differences observed effect in nature? Herpetol. J. 5: 204±205.
between ponds approach a magnitude generally con- Berven, K. A. (1987). The heritable basis of variation in larval
developmental patterns within populations of the wood frog
sidered as diagnostic between brown frog species (e.g. (Rana sylvatica). Evolution 41: 1088±1097.
Barbadillo et al., 1999). Together with the high Blair, J. & Wassersug, R. J. (2000). Variation in the pattern of
degree of damage of lingual papillae detected by the predator-induced damage to tadpole tails. Copeia 2000:
SEM studies, they are in the range of the differences 390±401.
found by Vences, Kupfer et al. (1998) among Pyre- Blouin, M. S. (1992). Genetic correlations among morphometric
nean R. temporaria populations. Thereby, our results traits and rates of growth and differentiation in the green tree
do not support the hypothesis that more than one frog, Hyla cinerea. Evolution 46: 735±744.
Brand, M. & Grossenbacher, K. (1979). Untersuchungen zur
temporaria-like Rana species occurs in this mountain Entwicklungsgeschwindigkeit der Larven von Triturus a. alpestris
chain. (Laurenti, 1768), Bufo b. bufo (Linnaeus, 1758) und Rana t.
A keratodont formula with only two interrupted temporaria (Linnaeus, 1758) aus Populationen verschiedener
upper rows seems to be typical for montane R. tempor- HoÈhenstufen in den Schweizer Alpen. PhD Dissertation, Univer-
aria populations (Sperling, Vences & BoÈhme, 1996; this sity of Bern.
study). These populations furthermore differ by accel- Bresler, J. & Bragg, A. N. (1954). Variations in the rows of labial
teeth in tadpoles. Copeia 1954: 255±257.
erated growth rates from lowland populations (Brand
Calef, G. W. (1973). Natural mortality of tadpoles in a population
& Grossenbacher, 1979) as is also found in other of Rana aurora. Ecology 54: 741±758.
anurans (Gollmann & Gollmann, 1996). Growth rate is Cooke, A. S. (1976). Neutral red dye as a marker for tadpoles. Br.
also higher in R. temporaria populations from northern J. Herpetol. 5: 701±705.
latitudes as compared to those from more southern Crump, M. L. (1989). Life history consequences of feeding versus
sites (MerilaÈ et al., 2000). Relative hindlimb length of non-feeding in a facultatively non-feeding toad larva. Oecologia
metamorphs may differ relative to size at metamor- (Berl.) 78: 486±489.
Denver, R. J. (1997). Proximate mechanisms of phenotypic plasti-
phosis in frogs, caused by heterochronic effects city in amphibian metamorphosis. Am. Zool. 37: 172±184.
(Emerson, Travis & Blouin, 1988). The variation in Denver, R. J., Mirhadi, N. & Phillips, M. (1998). Adaptive
larval growth history detected in the present study plasticity in amphibian metamorphosis: response of Scaphiopus
certainly in¯uences life history and survival of meta- hammondi tadpoles to habitat dessication. Ecology 79:
morphs. Additionally, its magnitude also makes an 1859±1872.
in¯uence on adult morphology conceivable. If such Dubois, A. (1995). Keratodont formula in anuran tadpoles: propo-
in¯uences were demonstrated by further studies, they sals for a standardization. J. Zoo Syst. Evol. Res. 33: I±XV.
Dutta, S. K. & Mohanty-Hejmadi, P. (1984). Ontogeny of teeth
may partly explain the extraordinary variability of row structure in Rana tigerina tadpoles. J. Bombay nat. Hist.
external morphology, both within populations and Soc. 80: 517±528.
between high- and low-altitude populations of R. tem- Emerson, S. B., Travis, J. & Blouin, M. (1988). Evaluating a
poraria. hypothesis about heterochrony: larval life-history traits and
juvenile hind-limb morphology in Hyla crucifer. Evolution 42:
Gollmann, B. & Gollmann, G. (1996). Geographic variation of
larval traits in the Australian frog Geocrinia victoriana. Herpe-
tologica 52: 181±187.
We are grateful to the Servicio Provincial de Agricultura Gosner, K. (1960). A simpli®ed table for staging anuran embryos
y Medio Ambiente of the DiputacioÂn de AragoÂn for and larvae with notes on their identi®cation. Herpetologica 16:
research and collecting authorizations. Antonio Palanca 183±190.
(Vigo) organized the ®eld course during which data Grillitsch, B. & Grillitsch, H. (1989). Teratological and ontoge-
collection took place. Carlos AcunÄa, RubeÂn Fraga, netic alterations to external oral structure in some anuran
larvae (Amphibia: Anura: Bufonidae, Ranidae). In Trends in
Xoan Santos, Sara Varela and AndreÂs Villaverde as- vertebrate morphology, Proc. 2nd Int. Symp. Verteb. Morphol.,
sisted in the ®eld. JesuÂs Mendez (Vigo) provided Vienna, 1986. Splechtna, H. & Hilgers, H. (Eds). Fortschri.
important help with the SEM analyses. Luis Villar Zool. 35: 276±281.
(Jaca) helped with determination of plants. Thanks also Gromko, M. H., Mason, F. S. & Smith-Gill, S. J. (1973). Analysis
to J. Kosuch and M. Veith (Mainz) who helped with the of the crowding effect in Rana pipiens tadpoles. J. exp. Zool.
genetic analyses, D. Tarkhnishvili (Bonn) for many 186: 63±72.
helpful discussions, and two anonymous referees and Harkey, G. A. & Semlitsch, R. D. (1988). Effects of temperature
on growth, development, and color polymorphism in the
Angus Carpenter (Norwich) for useful comments on the ornate chorus frog Pseudacris ornata. Copeia 1988: 1001±1007.
162 M. Vences et al.

Hillis, D. M. (1982). Morphological differentiation and adapta- PiqueÂ, N., LoÂpez, A., Vieites, D. R., Nieto-RomaÂn, S. & Vences,
tion of the larvae of Rana berlandieri and Rana sphenocephala M. (2000). Tadpole density, microhabitat use and daily activity
(Rana pipiens complex) in sympatry. Copeia 1982: 168±174. cycles in a Pyrenean population of Rana temporaria. Bol. Asoc.
Inger, R. F. (1992). A bimodal feeding system in a stream-dwelling Herpetol. Esp 11: 28±31.
larva of Rhacophorus from Borneo. Copeia 1992: 887±890. Potthoff, T. L. & Lynch, J. D. (1986). Interpopulation variability
Jennings, R. D. & Scott, N. J. Jr (1993). Ecologically correlated in mouthparts in Scaphiopus bombifrons in Nebraska (Am-
morphological variation in tadpoles of the leopard frog, Rana phibia: Pelobatidae). Prairie Nat. 18: 15±22.
chiricahuensis. J. Herpetol. 27: 285±293. Reques, R. & Tejedo, M. (1995). Negative correlation between
Kupferberg, S. J. (1997). The role of larval diet in anuran length of larval period and metamorphic size in natural popula-
metamorphosis. Am. Zool. 37: 146±159. tions of natterjack toads (Bufo calamita). J. Herpetol. 29:
Loman, J. (1999). Early metamorphosis in common frog Rana 311±314.
temporaria tadpoles at risk of drying: an experimental demon- Rowe, C. L., Kinney, O. M. & Congdon, J. D. (1998). Oral
stration. Amphib.-Reptilia 20: 421±430. deformities in tadpoles of the bullfrog (Rana catesbeiana)
McCollum, S. A. & Leimberger, J. D. (1997). Predator-induced caused by conditions in a polluted habitat. Copeia 1998:
morphological changes in an amphibian: predation by dragon- 244±246.
¯ies affects tadpole shape and color. Oecologia (Berl.) 109: Serra-Cobo, J. (1993). DescripcioÂn de una nueva especie
615±621. europea de rana parda (Amphibia, Anura, Ranidae). Alytes
MerilaÈ, J., Laurila, A., Laugen, A. T., RaÈsaÈnen, K. & Pahkala, M. 11: 1±15.
(2000). Plasticity in age and size at metamorphosis in Rana Sperling, P., Vences, M. & BoÈhme, W. (1996). VorlaÈu®ge Bemer-
temporaria ± comparison of high and low latitude populations. kungen zum taxonomischen Status von Rana temporaria hon-
Ecography 23: 457±465. norati Heron-Royer, 1881. Salamandra 32: 99±112.
Miranda, L. A. & PisanoÂ, A. (1993). Efecto de la densidad Tubbs, L. O. E., Stevens, R., Wells, M. & Altig, R. (1993).
poblacional en larvas de Bufo arenarum producido a traves de Ontogeny of the oral apparatus of the tadpole of Bufo amer-
senÄales visuales. Alytes 11: 64±76. icanus. Amphib.-Reptilia 14: 333±340.
Newman, R. A. (1989). Developmental plasticity of Scaphiopus Vences, M., Kosuch, J., LoÈtters, S., Widmer, A., KoÈhler, J.,
couchii tadpoles in an unpredictable environment. Ecology 70: Jungfer, K. H. & Veith, M. (2000). Phylogeny and classi®cation
1775±1787. of poison frogs (Amphibia: Dendrobatidae), based on mito-
Nishihara, A. (1996). Effect of density on growth of head size in chondrial 16S and 12S ribosomal RNA gene sequences. Mol.
larvae of the salamander Hynobius retardatus. Copeia 1996: Phylogenet Evol. 15: 34±40.
478±483. Vences, M., Kupfer, A., Llorente, G. A., Montori, A. & Carre-
Pearman, P. B. (1993). Effects of habitat size on tadpole popula- tero, M. A. (1998). Description of the larval stages of the
tions. Ecology 74: 1982±1991. Pyrenean frog, Rana pyrenaica Serra-Cobo, 1993 (Amphibia:
Petersen, C. G. J. (1896). The yearly immigration of young plaice Ranidae). Boll. Mus. reg. Sci. nat. Torino 15: 1±23.
into Limfjord from the German sea. Report of the Danish Viertel, B. (1982). The oral cavities of Central European anuran
Biological Station 6. Copenhagen. larvae (Amphibia). Morphology, ontogenesis and generic diag-
Petranka, J. W. (1995). Interference competition in tadpoles: are nosis. Amphib.-Reptilia 4: 327±360.
multiple agents involved? Herpetol. J. 5: 206±207. Viertel, B. & Richter, S. (1999). Anatomy: viscera and endocrines.
Pfennig, D. W. (1992). Polyphenism in spadefoot toad tadpoles as In Tadpoles. The biology of anuran larvae: 92±148. McDiarmid,
a locally adjusted evolutionary stable strategy. Evolution 46: W. W. & Altig, R. (Eds). Chicago: The University of Chicago
1408±1420. Press.
Pfennig, D. W. & Collins, J. P. (1993). Kinship affects morphogen- Wilbur, H. M. & Semlitsch, R. D. (1990). Ecological consequences
esis in cannibalistic salamanders. Nature (Lond.) 362: 836±838. of tail injury in Rana tadpoles. Copeia 1990: 18±24.