Field collection
We studied the F. limnocharis population at the Ankang Branch Farm of National Taiwan University, Taipei, Taiwan (24° 57’ 36’’N, 121° 31’ 24’’E). The rice fields,
consisting of sixteen 40m-by-40m paddies, have been cultivated for nearly 50 years.
There are two annual crops, the spring crop experiences cooler and more variable air temperature than the summer crop (see Results section), based on the information collected at a nearby weather station (data retrieved from the Data Bank for Atmospheric Research in Taiwan; http://dbar.as.ntu.edu.tw/). To determine the
relationship between air and water temperature, we recorded water temperature at three random locations in the rice fields in 2007 with water temperature sensors (HOBO® Pendant™, Onset Computer Corporation, MA, USA).
We randomly collected 40 F. limnocharis tadpoles at the developmental stage of Gosner 25 (Gosner, 1960) from several rice paddies in the farm on 07-May-2006 during the 1st rice crop (spring cohort, hereafter) and on 20-September-2006 during the 2nd rice crop (summer cohort, hereafter), and transported them immediately to the laboratory. Tadpoles remained at room temperature before the experiments started within 24 hours of collection.
Experimental design
We used a 2×2 factorial design to examine the effects of cohort (spring vs. summer) and temperature (low vs. high) on larval growth and development. The tadpoles were
housed in two identical environmental chambers (Chang Kuang Company, Taipei, Taiwan) with a 12:12 hour light and dark cycle. Daylight hours were similar between spring (12 hr. 56 min.) and summer (12 hr. 17 min.) months in northern Taiwan (t-test, t
= 1.20, df = 5, P = 0.29). We set chamber at 22°C (LT) and 29°C (HT) to represent the spring and summer temperature, respectively. Both temperatures were within the natural temperature range at the study site (see Results section). The four treatments were denoted as LT-spring, LT-summer, HT-spring, HT-summer. Thirty-two healthy tadpoles from each cohort were selected and randomly assigned to temperature treatments, 16 tadpoles each. Tadpoles were raised individually in opaque plastic containers (W×L×H:
7.5×11.5×4 cm) containing 250ml tap water, and provided boiled spinach ad libitum.
We measured weight and body dimensions and determined Gosner stage for each individual weekly until they reached metamorphosis (Gosner 42). We weighed
individuals with a precision weighing balance (accuracy: 0.1 mg, Sartorius, Germany).
We used the image analysis software SigmaScan Pro® ver. 5.0.0 (Jandel Scientific, USA) to determine body dimensions including snout-vent length and tail length. As the
forelimbs of tadpoles emerged, i.e., at the onset of metamorphosis (Gosner 42), we recorded the body weight (metamorphic weight, hereafter), and calculated the number of days elapsed between Gosner 25 and Gosner 42 (metamorphic age, hereafter). We calculated body growth rates as the difference in body weights between Gosner 25 and 42 divided by the metamorphic age. The metamorphosing tadpoles were placed in a large container with shallow water until their tails disappeared (Gosner 46), and then released back to the farm. Generally, tadpoles achieved maximum body weights at Gosner 40-41, then ceased to grow and lost body weight until metamorphosis was completed (Adolph, 1931). We measured tadpole body weights daily when tadpoles reached Gosner 40 and onward to find their maximum weights. We calculated the weight loss before entering metamorphosis as the difference between the maximum weight and the body weight at Gosner 42 (metamorphic weight). We divided the weight loss by maximum weight to obtain the percentage of body weight loss during the pre-metamorphosis period. We also calculated the relative tail length at metamorphosis by dividing tail length by total length to provide information on allometric changes in the tadpoles.
Data analysis
We compared the monthly air temperature (average of daily temperature) and its
variability (coefficient of variation per season per year) between spring (March-June) and summer (August-October) crops at the study site from 1997 to 2007 using t-tests.
We also used a paired t-test to examine the difference in daily water and air temperatures in spring and summer 2007.
We started each experimental treatment with 16 tadpoles. However, several tadpoles died, and several had not entered G42 when we had to stop the experiment for logistic reasons. We excluded those tadpoles from analyses. Sample sizes were 7, 9, 11, 14 for the treatment of 22°C-spring, 22°C-summer, 29°C-spring and 29°C-summer, respectively. We used fixed factor (temperature and cohort) two-way ANOVAs to test for differences among treatments for variables including body dimensions, metamorphic weight, metamorphic age, body growth rate, and percentage of weight loss before
entering metamorphosis, and Scheffé’s tests to perform post hoc paired comparisons.
Among all variables, only the metamorphic age data set didn’t conform to a normal
distribution (Shapiro-Wilk test for normality, P <0.001) and was treated with a square root transformation before further analysis. All measured variables were reported in mean±1se. All statistical analyses were conducted using SYSTAT 11 (Systat Software Inc., CA, USA).
Results
Environmental temperature
Spring crops experienced cooler and more variable air temperature than summer crops.
(spring 22.81±0.16°C, summer 25.89±0.19°C, t-test, t = 16.70, df = 9, P < 0.001; range of daily temperature, spring 3.3-37.3°C, summer 15.2-39.3°C; coefficient of variation, spring 21.67±0.57%, summer 14.96±0.43%, t = 5.68, df = 9, P < 0.001). Air
temperature reflected water temperature well, the two were not significantly different in 2007 (paired t-test, t = 1.52, df = 7, P = 0.17). Spring crops experienced cooler water temperature than summer crops in 2007 (21.55±1.03°C and 25.37±0.60°C,
respectively).
Metamorphic weight
There was a significant interaction between temperature and cohort in metamorphic weight (Two-way ANOVA, cohort x temperature interaction, F1, 37 = 16.55, P < 0.001, Table 2.1). Spring tadpoles were heavier than summer tadpoles at low, but not high temperature (Fig. 2.1a).
Metamorphic age
There was a significant interaction between temperature and cohort in metamorphic age (Two-way ANOVA, cohort x temperature interaction, F1, 37 = 13.76, P < 0.001, Table 2.1). Summer tadpoles reached metamorphosis earlier (lower metamorphic age) than spring ones at high, but not low temperature (Fig. 2.1b).
Growth rate
There was a significant interaction between temperature and cohort in body growth rate from G25 to G42 (Two-Way ANOVA, cohort x temperature interaction, F1, 37 = 42.32, P
< 0.001, Table 2.1). Summer tadpoles had greater growth rate than spring ones at high temperature, whereas spring tadpoles had greater growth rate than summer ones at low temperature (Fig. 2.1c).
Body weight loss
There was a significant interaction between temperature and cohort (Two-way ANOVA, F1, 36 = 23.77, P < 0.001, Table 2.1) in the percentage of body weight lost during
pre-metamorphosis. Spring tadpoles lost less body weight at low temperature than high temperature, while summer tadpoles lost less body weight at high temperature than low temperature (Fig. 2.1d).
Metamorphic body dimensions
There were significant interactions between temperature and cohort in total length (Two-way ANOVA, cohort x temperature interaction, F1,37 = 10.76, P = 0.002, Table 2.1), relative tail length (F1,37 = 13.64, P = 0.001, Table 2.1). Spring tadpoles were longer than summer tadpoles in total length at low, but not high temperature at metamorphosis (Fig. 2.1e). In addition, spring, but not summer, tadpoles showed differential allometric changes when raised under different temperatures. Spring tadpoles had shorter relative tail length (Fig. 2.1f) at high, but not low temperature.
Discussion
The two cohorts of tadpoles of the Indian Rice Frog, F. limnocharis, encounter different environmental temperatures. The spring cohort faces low and variable temperatures, the summer cohort high and less variable ones. The results of our common garden
experiments showed that both spring and summer tadpoles had greater body growth rates (Fig. 2.1c) and less weight loss before metamorphosis (Fig. 2.1d) when raised under their respective field temperature. It suggests that the tadpoles had better assimilation (greater body growth rates) and transformation (less weight loss)
efficiencies, and thus potentially higher fitness in their respective field temperature than otherwise (Benavides et al., 2005).
Furthermore, we found the two cohorts use different life-history strategies to cope with their thermal environments. The spring tadpoles responded to low temperature with high body weights at metamorphosis, while the summer tadpoles did not have such a response (Fig. 2.1a). On the other hand, while both spring and summer tadpoles responded to high temperature with accelerated developmental rates, summer tadpoles developed significantly faster than the spring ones. Thus, the summer tadpoles raised at high, but not low, temperature had a significantly shorter larval period than the spring ones (Fig. 2.1b). Either getting out of the water early or metamorphosing at a larger size
increases fitness (Smith, 1983; Newman, 1987; Smith, 1987). That is, the spring and summer cohorts employed distinct life-history strategies to increase their fitness: the former used a ''size'' strategy, the latter a ''rate'' strategy.
The temperature experienced during the larval period has a crucial influence on amphibian metamorphosis. High temperatures not only facilitate metabolic rate and assimilation efficiency, but also stimulate endocrine secretions like thyroxin which are related to the hormonal regulation of metamorphosis (Smith-Gill & Berven, 1979; Beck
& Congdon, 2000; Benavides et al., 2005; Rose, 2005; Castañeda et al., 2006). Previous studies found that larval anurans raised at low temperatures would metamorphose at large sizes with long larval periods, while at higher temperatures they would
metamorphose at small sizes with short larval periods (Smith-Gill & Berven, 1979;
Harkey & Semlitsch, 1988; Morand, Joly & Grolet, 1997; Alvarez & Nicieza, 2002b;
Loman, 2002; Laugen et al., 2003; Watkins & Vraspir, 2006; Castañeda et al., 2006).
For example, a study of R. temporaria in Sweden showed that a 3°C difference in mean maximum temperature resulted in a 30% increase in growth rate during Gosner 25~42 (Stahlberg et al., 2001). Our results conformed to the previous studies with an extra twist. Both spring and summer tadpoles responded to high temperature (29°C) with accelerated developmental rates to reach metamorphosis, but summer tadpoles increased
developmental rates by >300%, compared to < 200% increase in spring tadpoles (Fig.
2.1b). Due to the much accelerated developmental rates, the body growth rates of summer tadpoles raised under high temperature were four times faster than those under low temperature (22°C). The spring cohorts did not have such a dramatic differential response (Fig. 2.1c).
Why do the two cohorts use different strategies? We found spring tadpoles experienced greater temperature variations than summer tadpoles. Fluctuating thermal conditions are greater environmental stress than stable ones for ectothermal animals.
Large individuals are equipped with better physiological and locomotive capabilities and are therefore more capable of handling unpredictable conditions than small ones (Beck & Congdon, 2000; Morey & Reznick, 2001; Altwegg & Reyer, 2003; Wells, 2007). In addition, individuals in the spring cohort could potentially start breeding (Smith, 1987; Berven, 1988) in the same year, and large metamorphs would have advantages over small ones during breeding. (Alexander et al., 1979) showed that spring F. limnocharis tadpoles could assume breeding activities in August or September of the same year. Tadpoles that metamorphose at larger sizes have the potential to increase their lifetime reproductive success.
The summer condition was different. Summer tadpoles faced temperatures as high
as 39°C. The fast developmental rate could help the summer tadpoles cope with
growing season constraints and predation pressure in the summer and fall. The temporal limitation imposed by the ending of the growing season could accelerate anuran larval growth rate (Berven et al., 1979; Berven & Gill, 1983; Blouin & Brown, 2000;
Stahlberg et al., 2001; Laugen et al., 2003). The summer cohort has a shorter growing season than the spring cohort in Taiwan because the temperature drops and food availability on land declines as fall approaches and the summer rice crop is harvested.
Early metamorphs of the summer cohort could spend a longer time preparing for winter.
In addition, predation pressure is likely higher in summer than spring rice paddies. Rice paddies are ephemeral water bodies, and do not support predatory fish. Major avian predators, such as little egrets, Egretta garzetta, are mainly migrants, and do not arrive Taiwan until late summer or fall. Invertebrate predators such as dragonfly nymphs only occur in rice fields during August in northern Taiwan (Personal observation by S.-H.
Kuan and communication with L.-J. Wang). With fast developmental rates, the summer cohort could have a better chance of escaping the risks of predation.
It was interesting to observe that spring, but not summer, tadpoles had differential allometric growth patterns when raised under different temperatures. Spring tadpoles raised under high temperatures had relatively shorter tails than those from the other
three treatments. Such a morphology often translates into poor swimming ability (Stahlberg et al., 2001). Under predatory pressure, tadpoles often change their morphological and/or behavioral traits, including relative tail lengths and depths to reduce the risk of predation (reviewed in Benard, 2004). No study thus far has examined responses of F. limnocharis to the risk of predation, and the consequences of the
responses.
The different life-history strategies used by the spring and summer cohorts could be attributed to either maternal effect, temporal genetic divergence, or both (Alvarez &
Nicieza, 2002b; Benavides et al., 2005; Lesbarreres et al., 2007). These factors will need to be investigated further. Particularly, different temperature regimes between the two rice crops might have resulted in temporally distinct cohorts of F. limnocharis. A population genetics study of F. limnocharis indicated high genetic differentiation on a local scale; 84% of genetic variance came from individuals within subpopulations (Guan, 2005). A study is underway to investigate whether F. limnocharis larvae living in the rice fields have temporally diverged within the same habitat.
Table 2.1 The two-way ANOVAs, cohort × temperature, examine the (a) metamorphic weight, (b) metamorphic age, (c) body growth rate, (d) percentage of body weight loss, (e) total length, and (f) relative tail length at metamorphosis (Gosner 42) of spring and summer Indian Rice frog tadpoles raised under two different water temperatures.
Source Sum of
(a) (b)
(c) (d)
(e) (f)
Fig. 2.1 The (a) metamorphic weight, (b) metamorphic age, (c) body growth rate, (d) percentage of body weight lost, (e) total length, (f) relative tail length at metamorphosis (Gosner 42) of spring and summer F. limnocharis tadpoles (spring, white bars; summer, grey bars) raised under two temperatures. Value were given in MEAN±1SE. Different letters on the bars indicated significant difference (P < 0.05) based on post hoc paired comparisons using Scheffé’s tests. Sample sizes were 7, 9, 11, 14 for the treatment of 22°C-spring, 22°C-summer, 29°C-spring and 29°C-summer, respectively.