doi: 10.1111/jeb.12653
Geographic body size variation in the periodical cicadas Magicicada: implications for life cycle divergence and local adaptation T. KOYAMA*, H. ITO†, S. KAKISHIMA†, J. YOSHIMURA†‡§, J. R. COOLEY¶, C. SIMON¶ & T. SOTA* *Department of Zoology, Graduate School of Science, Kyoto University, Kyoto, Japan †Graduate School of Science and Technology, Shizuoka University, Hamamatsu, Japan ‡Department of Environmental and Forest Biology, College of Environmental Science and Forestry, State University of New York, Syracuse, NY, USA §Marine Biosystems Research Center, Chiba University, Uchiura, Kamogawa, Chiba, Japan ¶Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT, USA
Keywords:
Abstract
clinal variation; converse Bergmann cline; geographic variation; sexual dimorphism.
Seven species in three species groups (Decim, Cassini and Decula) of periodical cicadas (Magicicada) occupy a wide latitudinal range in the eastern United States. To clarify how adult body size, a key trait affecting fitness, varies geographically with climate conditions and life cycle, we analysed the relationships of population mean head width to geographic variables (latitude, longitude, altitude), habitat annual mean temperature (AMT), life cycle and species differences. Within species, body size was larger in females than males and decreased with increasing latitude (and decreasing habitat AMT), following the converse Bergmann’s rule. For the pair of recently diverged 13- and 17-year species in each group, 13-year cicadas were equal in size or slightly smaller on average than their 17-year counterparts despite their shorter developmental time. This fact suggests that, under the same climatic conditions, 17-year cicadas have lowered growth rates compared to their 13-years counterparts, allowing 13-year cicadas with faster growth rates to achieve body sizes equivalent to those of their 17-year counterparts at the same locations. However, in the Decim group, which includes two 13-year species, the more southerly, anciently diverged 13-year species (Magicicada tredecim) was characterized by a larger body size than the other, more northerly 13- and 17-year species, suggesting that local adaptation in warmer habitats may ultimately lead to evolution of larger body sizes. Our results demonstrate how geographic clines in body size may be maintained in sister species possessing different life cycles.
Introduction Many animal species with broad geographic ranges show clinal geographical variation in body size following Bergmann’s rule or the converse of Bergmann’s rule (Blanckenhorn & Demont, 2004). Insect species with relatively longer generation times typically show converse Bergmann clines (Blanckenhorn & Demont, Correspondence: Teiji Sota, Department of Zoology, Graduate School of Science, Kyoto University, Sakyo, Kyoto 606-8502, Japan. Tel.: +81 75 753 4078; fax: +81 75 753 4101; e-mail: [email protected]. kyoto-u.ac.jp
1270
2004), a likely result of life-history optimization to growing season length (Masaki, 1967; Roff, 1980). Such clines are expected to be modified when a species occupying a wide geographic range shows a drastic life cycle shift along a climatic gradient. For example, a shift from bivoltine to univoltine life cycle produces a saw-tooth cline in a cricket due to the abrupt change in the length of the juvenile period between the life cycles (Masaki, 1972; Roff, 1980). In addition, while allopatric sister species are expected to show similar body size clines when they have the same life cycle as has been demonstrated for intraspecific cases in Drosophila subobscura
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Body size variation in periodical cicadas
(e.g. Huey et al., 2000; Gilchrist et al., 2001), the body size cline may vary among closely related species as a result of a compound effect of colonization history, life cycle shift and interspecific interactions. For example, closely related species of carabid beetles with overlapping ranges in Japanese islands show similar slopes of body size clines but have different body sizes in sympatry (Sota et al., 2000a,b); their body size differences may have been affected by selection for avoiding hybridization (Okuzaki et al., 2010). To understand the complex evolution of body size clines, we conducted a detailed analysis of geographic variation in body size for a clade of cicadas occupying a wide geographic range. Periodical cicadas of the genus Magicicada occur only in the eastern United States and are renowned for the longest juvenile periods among insects (13 or 17 years). They consist of three species groups (Decim, Cassini and Decula; Fig. 1a), and each group has morphologically similar 13- and 17-year life cycle forms (Alexander & Moore, 1962; Marshall & Cooley, 2000). Seventeenyear cicadas live generally north of 13-year cicadas, with a narrow overlapping zone between their ranges (Cooley et al., 2009). The overall latitudinal range of Magicicada is wide, from 30.4467°N (LA) to 43.0669°N (WI, NY), with the northern limit of 17-year cicadas being approximately 3° farther north than limit of 13year cicadas. Of the three species groups, the Decim group includes two 13-year species, Magicicada tredecim and Magicicada neotredecim, and one 17-year species Magicicada septendecim. Both the Cassini and Decula groups consist of a pair of 13-/17-year species, Magicicada tredecassini/Magicicada cassini and Magicicada tredecula/ Magicicada septendecula, respectively. A recent molecular phylogenetic analysis showed that 13- and 17-year cicadas diverged independently within each species group (Sota et al., 2013). Of the currently recognized seven 13- and 17-year species, M. tredecim was estimated to have diverged from other Decim species 0.5 million years ago (Fig. 1a). Except for this deeper divergence, the remaining three pairs of 13- and 17-year species in the three species groups appear to have diverged no more than 23 000 years ago (Sota et al., 2013). The difference in life cycle length translates into a potential difference in fitness, because a 13-year life cycle has a 1.3 times greater intrinsic rate of population increase than a 17-year life cycle if all else is equal. However, extension of the juvenile period may affect fitness differently through its promotion of increased adult body size. Larger body size may be associated with greater fecundity in females (Karban, 1997) and increased mating success in males (Lloyd & Dybas, 1966; Karban, 1983). If adult body size does increase with an increase in life cycle length, 13-year cicadas may have smaller bodies and lower reproductive capacities than 17-year cicadas under the same habitat conditions, as is assumed in recent models of cicada evolution (Tanaka et al., 2009; Yoshimura et al., 2009). The correlation between body
body size [life cycle, y] species
Species group
(a)
2.906 2.907
Decula
2.906 2.903 Cassini
2.885 2.890
2.911 Decim 2.931
5
4
2 3 Age, mya
1
1271
[17] M. septendecula [13] M. tredecula [17] M. cassini
2.880
[13] M. tredecassini
2.928
[17] M. septendecim
2.918
[13] M. neotredecim
2.943
[13] M. tredecim
0
(b) 45°
40°
Cycle 17-y 13-y
35°
30°
95°
90°
85°
80°
75°
70°
Fig. 1 (a) Phylogeny and divergence times of Magicicada species (modified from Sota et al., 2013). Numerals at tree tips are mean body size (log10-transformed head width in 10-lm unit) of each species adjusted for sex and habitat annual mean temperature (AMT) (Table 2); numerals at tree nodes are reconstructed ancestral body sizes. (b) Sampling locations. Open and closed circles indicate 13- and 17-year cicadas, respectively.
size and life cycle length will hold if 13- and 17-year cicadas have the same annual nymphal growth rates under the same habitat conditions (Fig. 2 line 1). However, it is suggested that 17-year cicadas have a developmental arrest during the early juvenile period (White & Lloyd, 1975). If the additional 4 years of 17-year cicadas under the habitat conditions sufficient for 13-year life cycle are effectively a period of developmental arrest that adjusts the life cycle length to 17 years, 17-year cicadas would be expected to attain body sizes equivalent to those of 13-year cicadas (Fig. 2 line 2). Thus, the relatively recent putative life cycle shifts may have resulted in little difference in body size between the two life cycle lengths under the same habitat conditions, and the 13-year life cycle may have attained higher fitness than 17-year life cycle simply due to the shortened generation time. The relationship between life cycle length and body size has not been established by empirical studies except indirectly by studies of wing morphometric characters
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1272
T. KOYAMA ET AL.
Brood XIX (2011) and fixed them in 99% ethanol (Fig. 1b; Table S1). Note that specific brood numbers in roman numerals are assigned for cohorts of Magicicada that will emerge as adults at 13- or 17-year intervals (Marlatt, 1907). The specimens were stored at 30 to 20 °C. To estimate body size, we selected head width as measured by the distance between the outer edges of the eyes, because we determined that this measure could be taken reliably for every adult specimen. We measured the head width using digital vernier callipers for 4–65 samples from each location for a total of 4719 samples (Table S1). Explanatory variables
(Simon, 1983, 1992), which did not account for variation in relation to temperature/latitude. In this study, we investigated the adult body sizes of all seven Magicicada species throughout their geographic ranges to determine whether the species show clinal body size variation with geographic and climatic variables, and whether the body sizes of 13- and 17-year species within the same species group show any differences related to their life cycle differences as well as their history of divergence. Specifically, we tested whether the body sizes of 13- and 17-year cicadas at the same geographic location are similar to each other as expected if 17-year cicadas experience lower overall growth rates, as proposed by White & Lloyd (1975). We discuss the evolutionary processes and selective forces leading to the observed pattern of body size variation along a climatic gradient, between 13- and 17-year life cycles, and among species with different phylogenetic histories.
Potentially, a large number of geographic and climatic variables could be used in the analysis of insect body size variation across a geographic range. We initially tested latitude, longitude, altitude and multiple climatic variables included in WorldClim database (Hijmans et al., 2005; http://www.worldclim.org) as the variables to explain variation in body size. However, the WorldClim variables generally showed high correlations with one another and with latitude (Table S2). Such high correlations between explanatory variables caused the problem of multicollinearity and made the interpretation of statistical results difficult. Therefore, we decided to use geographic and climatic variables separately in different models. Further, we selected a single climatic variable, annual mean temperature (AMT) at each sampling location as the representative variable that reflects climatic condition. AMT data were obtained at a resolution of 30 s of latitude/longitude from WorldClim. We expect that AMT is related to forest productivity, on which the growth of cicada nymphs depends. Previous work by Cooley et al. (2013) showed that AMT was a strong predictor of 17-year cicada habitat suitability. We also have checked how AMT is related to ‘season length in day-degree unit’ measured by cumulative daily temperature above a developmental threshold temperature, which is often used as the explanatory climatic variable for clinal body size variation in insects. Because developmental threshold temperature for periodical cicadas is unknown, we obtained annual cumulative day-degrees above 0 °C using monthly mean temperatures for each geographic location. The season length thus obtained was highly correlated with AMT (n = 152 locations, r = 0.9926, P < 0.0001), and use of this variable instead of AMT yielded no substantial difference in statistical results.
Materials and methods
Statistical analyses
Fig. 2 Hypothetical growth curves for 13- and 17-year cicadas (body size is head width in logarithmic scale). Lines (1) represent constant growth curve shared by both 13- and 17-year cicadas; line (2) represents constant growth comparable to line 1 after a period of slow growth in early instars for 17-year cicadas.
Sampling and measurement of body size We collected 17-year cicadas of brood XIII (2007), XIV (2008), I (2012) and II (2013) and 13-year cicadas of
We conducted multiple regression analyses for the effects of sex, species group, species nested within species group, geographic variable (latitude, longitude, altitude) and AMT on body size (head width) using the
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1273
Body size variation in periodical cicadas
statistical package JMP version 10.0 (SAS Institute Inc., Cary, NC, USA). Head width data were normalized by log transformation, and mean values were calculated for each sex at each locality. In the multiple regression analysis, the dependent variable (mean head width) was weighted by sample size. Because latitude and AMT were strongly corrected with each other, we conducted analyses with geographic variables and AMT separately to avoid the problem of multicollinearity. In the model with geographic variables, all interaction terms with the geographic variables were initially incorporated and nonsignificant terms were removed in the final model. Following the multiple regression analysis with AMT, a Tukey HSD test was used to compare species’ mean body sizes (least square means). Also, we reconstructed body size at six ancestral nodes of the Magicicada phylogenetic tree from the species’ mean body sizes using ‘ace’ function of the package ‘ape’ (Paradis et al., 2004) in R version 2.15.2 (R Development Core Team, 2012). Comparison of nymphal instar compositions between 13- and 17-year cicadas To compare the developmental patterns during nymphal periods of 13- and 17-year cicadas, we analysed the instar composition data reported in White & Lloyd (1975) and Maier (1996). We tested the difference in instar composition between populations by G-test and the relationship between instar composition and habitat AMT by ordinal logistic regression. We used R version 2.15.2 for these statistical tests.
Results In the multiple regression models taking into account geographic or climatic factors, body size was negatively related to latitude and positively related to AMT (Table 1; Fig. 3). In the model with geographic vari-
Table 1 Multiple regression analysis for the effects of species group, species (nested in species group), sex, geographic variables and annual mean temperature (AMT) on body size (log10transformed head width). The effects of geographic variables and AMT were examined in separate models. Variable
d.f.
(a) Geographic variables Species group 2 Species 4 (species group) Sex 1 Latitude 1 Longitude 1 Altitude 1 Longitude*altitude 1 (b) AMT Species group 2 Species 4 (species group) Sex 1 AMT 1
F
P
Effect SE*
803.98 52.76
< 0.0001 < 0.0001
– –
605.10 60.20 0.26 1.33 10.95
< 0.0001 < 0.0001 0.6129 0.2495 0.0010
850.42 55.23
< 0.0001 < 0.0001
593.54 55.14
< 0.0001 < 0.0001
1.21E-02 2.04E-03 6.27E-05 5.79E-06 2.72E-06
4.93E-04 2.63E-04 1.24E-04 5.02E-06 8.22E-07
– – 1.22E-02 5.01E-04 2.25E-03 3.03E-04
*Effects for species group and species are not shown. See Table 2 for the result of multiple comparison of mean body sizes between species in the analysis on AMT (b).
ables (Table 1a), the effect of latitude was significant, as was the interaction effect of longitude and latitude. Having adjusted for these factors, sex, species group and species nested within species group showed consistently significant effects (Table 1). Females were larger than males in each species, demonstrating sexual size dimorphism (Table 2). The Decim group species were the largest, followed by species in the Decula group, and then those in the Cassini group (Table 2). In the Decim group, M. tredecim (13-year) was the largest, followed by M. septendecim (17-year) and then M. neotredecim (13-year). In the Decula group, body sizes did not differ between M. septendecula (17-year) and
Table 2 Mean head width (log10-transformed; in 10-lm unit) of each species and sex. Head width (least square mean SE) By species and sex† Species group
Species
Decim Decim Decim Decula Decula Cassini Cassini
Magicicada Magicicada Magicicada Magicicada Magicicada Magicicada Magicicada
tredecim septendecim neotredecim septendecula tredecula cassini tredecassini
Life cycle (years)
By species*
Male
13 17 13 17 13 17 13
2.943a 0.00137 2.928b 0.00131 2.918c 0.00145 2.907d 0.00181 2.906d 0.00233 2.890e 0.00126 2.880f 0.00123
2.931 2.919 2.909 2.889 2.888 2.877 2.866
Female
0.00172 0.00167 0.00185 0.00230 0.00331 0.00163 0.00161
2.954 2.937 2.926 2.926 2.922 2.902 2.893
0.00172 0.00169 0.00210 0.00266 0.00294 0.00162 0.00160
*Means adjusted for the effects of annual mean temperature (AMT) and sex following the multiple regression analysis given in Table 1b. Means followed by the same superscript letter do not differ significantly at P = 0.05 by Tukey HSD test. †Means by sex adjusted for the effect of AMT, showing that females are larger than males in each species. ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1274
T. KOYAMA ET AL.
M. tredecula (13-year). In the Cassini group, M. cassini (17-year) was significantly larger than M. tredecassini (13-year). In the analysis of nymphal instar composition data (Table 3), a 9-year-old 13-year population (Grand Tower) had a more advanced instar composition than a 9-year-old 17-year population at warmer habitat (Webber’s Falls) (G-test; G = 437.39, P < 2.2 9 10 16). However, instar composition did not differ significantly between 13- and 17-year populations 4 years before emergence (13-year-old Lewis Centre vs. 9-year-old Grand Tower populations: G-test, G = 4.14, P = 0.1264). (a) Decim male 2.98
(b)
M. septendecim M. neotredecim M. tredecim
2.93
2.88
2.88
2.83 7.0
11.0
15.0
19.0
(c) Cassini male
log10 Body size
2.98
2.98
2.93
2.88
2.88
15.0
19.0
(e) Decula male 2.98
15.0
19.0
M. cassini M. tredecassini
2.83 7.0
11.0
15.0
19.0
(f) Decula female
M. septendecula M. tredecula
2.98
2.93
M. septendecula M. tredecula
2.93
2.88
2.83 7.0
11.0
(d) Cassini female
M. cassini M. tredecassini
11.0
Our data reveal a complex pattern of sexually dimorphic body size variation related to climate, species (life cycle) and phylogeny. First, periodical cicadas showed a trend of increasing body size with decreasing
M. septendecim M. neotredecim M. tredecim
2.83 7.0
2.93
2.83 7.0
Discussion
Decim female
2.98
2.93
For the four 9th-year cohorts of 17-year cicadas, the effect of AMT on instar composition was significant (ordinal logistic regression: v2 = 185.54, P < 0.0001, b = 0.382), suggesting that nymphal development was slower in cooler habitats.
2.88
11.0
15.0
19.0
2.83 7.0
Annual mean temperature (°C)
11.0
15.0
19.0
Fig. 3 Relationship between body size (log10-transformed head width in 10lm unit) and annual mean temperature (AMT) by species for each sex in each species group. (a) Decim group, male; (b) Decim group, female; (c) Cassini group, male; (d) Cassini group, female; (e) Decula group, male; (f) Decula group, female. Bars indicate SE.
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Body size variation in periodical cicadas
1275
Table 3 Nymphal instar composition of 13- and 17-year cicada cohorts from six localities. Locality
Latitude (°N) Longitude (°W) Annual mean temperature (°C) Species Life cycle (year) Age at sampling (year) No. of nymphs 2nd instar 3rd instar 4th instar 5th instar Mean instar
Pilot Mound, IA*
Southington, CT†
Nashville, IN*
Webber’s Falls, OK*
Lewis Centre, OH*
Grand Tower, IL*
42.12 93.97 8.9
41.6 72.88 9.9
39.27 86.2 11.5
35.5 95.17 15.9
40.15 83.02 10.4
37.55 89.5 13.5
Magicicada cassini 17 9
Magicicada septendecim 17 9
Unknown
M. cassini
Unknown
Unknown
17 9
17 9
17 13
13 9
27 106 457 0 3.73
0 25 75 0 3.75
0 31 83 2 3.75
0 12 306 68 4.15
0 0 35 214 4.86
0 2 39 334 4.89
Source: *White & Lloyd (1975); †Maier (1996).
latitude (and increasing AMT), following the converse Bergmann’s rule (Table 1). This clinal pattern may indicate local adaptation to the length of growing season available for larval development; in southern (warmer) regions, annual time for development is longer, and large body size could be a result of selection for greater fecundity (Masaki, 1967; Roff, 1980; Karban, 1997). After removing climatic (AMT) effects, the anciently diverged 13-year M. tredecim showed the largest body size among the three Decim species. In the remaining six described species (excluding M. tredecim), 17-year species had similar or larger body sizes than their corresponding 13-year species after adjusting for the effect of AMT (Table 1; Fig. 3). This increase in body size is likely related to the 4-year extension of nymphal period of 17-year species compared to 13-year. However, the observed increase in body size (in log10 scale) was only 0.3–2.3% (Table 2). In contrast, based on the assumption that 17- and 13-year species grow at the same rate in log10 scale, 17-year cicadas are expected to be approximately 31% (4/13) larger than 17-year cicadas. Thus, 13-year cicadas must have faster growth rates on average than 17-year cicadas at the same AMT to achieve comparable body sizes. Both 13- and 17-year Magicicada species pass through five nymphal instars (White & Lloyd, 1975; Maier, 1996). Control of the nymphal period is unknown, but White & Lloyd (1975) postulated that 17-year cicadas spend longer nymphal periods in the early instars than do 13-year cicadas, implying that switching between 13- and 17-year cycles occurs by insertion/deletion of a 4-year period of developmental delay during the early juvenile period (Fig. 2). A full test of this hypothesis awaits age structure data throughout the entire juvenile periods for geographically close 13- and 17-year populations. For the
moment, our analysis of available data shows that nymphal instar composition 4 years before emergence in 13- and 17-year cicadas occurring at similar habitat temperatures did not differ whereas 9-year-old 13-year cicadas had a more advanced age structure than 9-year-old 17-year cicadas (Table 3). This suggests that the difference in development/growth rates between 13- and 17-year cicadas is attributable to the difference in early instars, specifically a slower development rate in early instars of 17-year species, as was postulated by White & Lloyd (1975). Thus, 17-year cicadas may experience a 4-year retardation of growth in early nymphal instars and thereafter grow as rapidly as 13-year cicadas to attain final body sizes similar to those of 13-year cicadas. In the Decim group, we found that the anciently diverged 13-year M. tredecim, unlike the recently diverged 13-year M. neotredecim, was larger than the 17-year M. septendecim (Table 1). This pattern may be the result of M. tredecim’s long-term adaptation to longer growing seasons given its generally southern distribution, while M. neotredecim individuals are small due to this species’ relatively recent divergence from M. septendecim (Marshall & Cooley, 2000; Simon et al., 2000; Sota et al., 2013) and the shorter growing seasons of the northern latitudes in which this species occurs. In conclusion, the body size differences between recently diverged 13- and 17-year species were negligible, while more anciently separated taxa in the Decim group showed notable differences, with M. tredecim larger than M. septendecim/neotredecim. Recent divergences between 13- and 17-year species in each species group (Sota et al., 2013) have been attributed to switching from the 17- to the 13-year cycle (e.g. Martin & Simon, 1988; Marshall & Cooley, 2000; Simon et al.,
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1276
T. KOYAMA ET AL.
2000; Cooley et al., 2001). The similarity in body size between recently diverged 13- and 17-year species in each species group suggests that along with the life cycle switch from 17- to 13-year cycles, there was a recovery of growth rates in early instars, which had been lowered in the 17-year cycle. Body size conservatism across a major life cycle shift appears to be unique to periodical cicadas. In insects, a drastic change in larval period length, such as from univoltine to bivoltine life cycles in crickets, is expected to result in a noticeable change in body size (Masaki, 1972; Roff, 1980; but see also Kivel€a et al., 2011 for counterexamples). Body size and associated dimensions may be related to other reproductive characters besides reproductive capacity such as male mating call frequency (Bennet-Clark & Young, 1994; but see Oberd€ orster & Grant, 2007); thus, a reduction in body size may reduce mating rate or likelihood of mating in addition to directly reducing reproductive capacity. Therefore, the body size conservatism in periodical cicadas during a life cycle switch may be adaptive for securing mating, although the resultant body size may not be optimal in terms of climatic adaptation.
Acknowledgments We thank H. Ikeda and D. Suzuki for sampling and W. Blanckenhorn for helpful comments on our manuscript. This study was supported by JSPS KAKENHI (no. 22255004, 22370010, 26257405 and 15H00420 to JY) and grants-in-aid for JSPS Fellows to HI and SK. CS and JRC received support from NSF DEB 09-55849.
References Alexander, R.D. & Moore, T.E. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species (Homoptera: Cicadidae, Magicicada). Univ. Mich. Mus. Zool. Misc. Pub. 121: 1–59. Bennet-Clark, H. & Young, D. 1994. The scaling of song frequency in cicadas. J. Exp. Biol. 191: 291–294. Blanckenhorn, W.U. & Demont, M. 2004. Bergmann and converse Bergmann latitudinal clines in arthropods: two ends of a continuum? Integr. Comp. Biol. 44: 413–424. Cooley, J.R., Simon, C., Marshall, D.C., Slon, K. & Ehrhardt, C. 2001. Allochronic speciation, secondary contact, and reproductive character displacement in periodical cicadas (Hemiptera: Magicicada spp.): genetic, morphological, and behavioural evidence. Mol. Ecol. 10: 661–671. Cooley, J.R., Kritsky, G., Edwards, M.J., Zyla, J.D., Marshall, D.C., Hill, K.B.R. et al. 2009. The distribution of periodical cicada brood X in 2004. Am. Entomol. 55: 106–112. Cooley, J.R., Marshall, D.C., Simon, C., Neckermann, M.L. & Bunker, G. 2013. At the limits: habitat stability modeling of northern 17-year periodical cicada extinctions (Hemiptera: Magicicada spp.). Glob. Ecol. Biogeogr. 22: 410–421. Gilchrist, G.W., Huey, R.B. & Serra, L. 2001. Rapid evolution of wing size clines in Drosophila subobscura. Genetica 112–113: 273–286.
Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G. & Jarvis, A. 2005. High resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25: 1965–1978. Huey, R.B., Gilchrist, G.W., Carlson, M.L., Berrigan, D. & Serra, L. 2000. Rapid evolution of a geographic cline in size in an introduced fly. Science 287: 308–309. Karban, R. 1983. Sexual selection, body size and sex-related mortality in the cidada Magiciada cassini. Am. Midl. Nat. 109: 324–330. Karban, R. 1997. Evolution of prolonged development: a life table analysis for periodical cicadas. Am. Nat. 150: 446–461. Kivel€a, S.M., V€alim€aki, P., Carrasco, D., M€aenp€a€a, M.I. & Oksanen, J. 2011. Latitudinal insect body size clines revisited: a critical evaluation of the saw-tooth model. J. Anim. Ecol. 80: 1184–1195. Lloyd, M. & Dybas, H.S. 1966. The periodical cicada problem. II. Evolution. Evolution 20: 466–505. Maier, C.T. 1996. Connecticut is awaiting return of the periodical cicada. Front. Plant Sci. 48: 4–6. Marlatt, C.L. 1907. The periodical cicada. US Dept. Agric. Bureau Entomol. Bull. 71: 1–183. Marshall, D.C. & Cooley, J.R. 2000. Reproductive character displacement and speciation in periodical cicadas, with description of a new species, 13-year Magicicada neotredecim. Evolution 54: 1313–1325. Martin, A.P. & Simon, C. 1988. Anomalous distribution of nuclear and mitochondrial DNA markers in periodical cicadas. Nature 336: 237–239. Masaki, S. 1967. Geographic variation and climatic adaptation in a field cricket (Orthoptera: Gryllidae). Evolution 21: 725– 741. Masaki, S. 1972. Climatic adaptation and photoperiodic response in the band-legged cricket. Evolution 26: 587–600. Oberd€ orster, U. & Grant, P.R. 2007. Acoustic adaptations of periodical cicadas (Hemiptera: Magicicada). Biol. J. Linn. Soc. 90: 15–24. Okuzaki, Y., Takami, Y. & Sota, T. 2010. Resource partitioning or reproductive isolation: the ecological role of body size differences among closely related species in sympatry. J. Anim. Ecol. 79: 383–392. Paradis, E., Claude, J. & Strimmer, K. 2004. APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20: 289–290. R Development Core Team 2012. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Roff, D. 1980. Optimizing development time in a seasonal environment: the ‘ups and downs’ of clinal variation. Oecologia 45: 202–208. Simon, C. 1983. Morphological differentiation in wing venation among broods of 13- and 17-year periodical cicadas. Evolution 37: 104–115. Simon, C. 1992. Discriminant analysis of year classes of periodical cicada based on wing morphometric data enhanced by molecular information. In: Ordinations in the Study of Morphology Evolution and Systematics of Insects: Applications and Quantitative Genetic Rationales (J.T. Sorensen & T.G. Footit, eds), pp. 309–322. Elsevier, Amsterdam, The Netherland. Simon, C., Tang, J., Dalwadi, S., Staley, G., Deniega, J. & Unnasch, T.R. 2000. Genetic evidence for assortative mating between 13-year cicadas and sympatric “17-year cicadas
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Body size variation in periodical cicadas
with 13-year life cycles” provides evidence for allochronic speciation. Evolution 54: 1326–1336. Sota, T., Takami, Y., Kubota, K. & Ishikawa, R. 2000a. Geographic variation in the body size of some Japanese Leptocarabus species (Coleoptera, Carabidae): the “toppleddomino pattern” in species along a geographic cline. Entomol. Sci. 3: 309–320. Sota, T., Takami, Y., Kubota, K., Ujiie, M. & Ishikawa, R. 2000b. Interspecific body size differentiation in species assemblage of the carabid subgenus Ohomopterus in Japan. Popul. Ecol. 42: 279–291. Sota, T., Yamamoto, S., Cooley, J.R., Hill, K.B.R., Simon, C. & Yoshimura, J. 2013. Independent divergence of 13- and 17-y life cycles among three periodical cicada lineages. Proc. Natl. Acad. Sci. USA 110: 6919–6924. Tanaka, Y., Yoshimura, J., Simon, C., Cooley, J.R. & Tainaka, K. 2009. Allee effect in the selection for prime-numbered cycles in periodical cicadas. Proc. Natl. Acad. Sci. USA 106: 8975–8979. White, J. & Lloyd, M. 1975. Growth-rates of 17-year and 13year periodical cicadas. Am. Midl. Nat. 94: 127–143.
1277
Yoshimura, J., Hayashi, T., Tanaka, Y., Tainaka, K. & Simon, C. 2009. Selection for prime-number intervals in a numerical model of periodical cicada evolution. Evolution 63: 288– 294.
Supporting information Additional Supporting Information may be found in the online version of this article: Table S1 Sampling locations of each species. Table S2 Correlation coefficients between latitude, annual mean temperature and other climate parameters for sampling locations in this study. Data deposited at Dryad: doi:10.5061/dryad.mn200 Received 8 February 2015; revised 24 February 2015; accepted 20 April 2015
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Geographic body size variation in the periodical cicadas Magicicada: implications for life cycle divergence and local adaptation T. KOYAMA*, H. ITO†, S. KAKISHIMA†, J. YOSHIMURA†‡§, J. R. COOLEY¶, C. SIMON¶ & T. SOTA* *Department of Zoology, Graduate School of Science, Kyoto University, Kyoto, Japan †Graduate School of Science and Technology, Shizuoka University, Hamamatsu, Japan ‡Department of Environmental and Forest Biology, College of Environmental Science and Forestry, State University of New York, Syracuse, NY, USA §Marine Biosystems Research Center, Chiba University, Uchiura, Kamogawa, Chiba, Japan ¶Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT, USA
Keywords:
Abstract
clinal variation; converse Bergmann cline; geographic variation; sexual dimorphism.
Seven species in three species groups (Decim, Cassini and Decula) of periodical cicadas (Magicicada) occupy a wide latitudinal range in the eastern United States. To clarify how adult body size, a key trait affecting fitness, varies geographically with climate conditions and life cycle, we analysed the relationships of population mean head width to geographic variables (latitude, longitude, altitude), habitat annual mean temperature (AMT), life cycle and species differences. Within species, body size was larger in females than males and decreased with increasing latitude (and decreasing habitat AMT), following the converse Bergmann’s rule. For the pair of recently diverged 13- and 17-year species in each group, 13-year cicadas were equal in size or slightly smaller on average than their 17-year counterparts despite their shorter developmental time. This fact suggests that, under the same climatic conditions, 17-year cicadas have lowered growth rates compared to their 13-years counterparts, allowing 13-year cicadas with faster growth rates to achieve body sizes equivalent to those of their 17-year counterparts at the same locations. However, in the Decim group, which includes two 13-year species, the more southerly, anciently diverged 13-year species (Magicicada tredecim) was characterized by a larger body size than the other, more northerly 13- and 17-year species, suggesting that local adaptation in warmer habitats may ultimately lead to evolution of larger body sizes. Our results demonstrate how geographic clines in body size may be maintained in sister species possessing different life cycles.
Introduction Many animal species with broad geographic ranges show clinal geographical variation in body size following Bergmann’s rule or the converse of Bergmann’s rule (Blanckenhorn & Demont, 2004). Insect species with relatively longer generation times typically show converse Bergmann clines (Blanckenhorn & Demont, Correspondence: Teiji Sota, Department of Zoology, Graduate School of Science, Kyoto University, Sakyo, Kyoto 606-8502, Japan. Tel.: +81 75 753 4078; fax: +81 75 753 4101; e-mail: [email protected]. kyoto-u.ac.jp
1270
2004), a likely result of life-history optimization to growing season length (Masaki, 1967; Roff, 1980). Such clines are expected to be modified when a species occupying a wide geographic range shows a drastic life cycle shift along a climatic gradient. For example, a shift from bivoltine to univoltine life cycle produces a saw-tooth cline in a cricket due to the abrupt change in the length of the juvenile period between the life cycles (Masaki, 1972; Roff, 1980). In addition, while allopatric sister species are expected to show similar body size clines when they have the same life cycle as has been demonstrated for intraspecific cases in Drosophila subobscura
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Body size variation in periodical cicadas
(e.g. Huey et al., 2000; Gilchrist et al., 2001), the body size cline may vary among closely related species as a result of a compound effect of colonization history, life cycle shift and interspecific interactions. For example, closely related species of carabid beetles with overlapping ranges in Japanese islands show similar slopes of body size clines but have different body sizes in sympatry (Sota et al., 2000a,b); their body size differences may have been affected by selection for avoiding hybridization (Okuzaki et al., 2010). To understand the complex evolution of body size clines, we conducted a detailed analysis of geographic variation in body size for a clade of cicadas occupying a wide geographic range. Periodical cicadas of the genus Magicicada occur only in the eastern United States and are renowned for the longest juvenile periods among insects (13 or 17 years). They consist of three species groups (Decim, Cassini and Decula; Fig. 1a), and each group has morphologically similar 13- and 17-year life cycle forms (Alexander & Moore, 1962; Marshall & Cooley, 2000). Seventeenyear cicadas live generally north of 13-year cicadas, with a narrow overlapping zone between their ranges (Cooley et al., 2009). The overall latitudinal range of Magicicada is wide, from 30.4467°N (LA) to 43.0669°N (WI, NY), with the northern limit of 17-year cicadas being approximately 3° farther north than limit of 13year cicadas. Of the three species groups, the Decim group includes two 13-year species, Magicicada tredecim and Magicicada neotredecim, and one 17-year species Magicicada septendecim. Both the Cassini and Decula groups consist of a pair of 13-/17-year species, Magicicada tredecassini/Magicicada cassini and Magicicada tredecula/ Magicicada septendecula, respectively. A recent molecular phylogenetic analysis showed that 13- and 17-year cicadas diverged independently within each species group (Sota et al., 2013). Of the currently recognized seven 13- and 17-year species, M. tredecim was estimated to have diverged from other Decim species 0.5 million years ago (Fig. 1a). Except for this deeper divergence, the remaining three pairs of 13- and 17-year species in the three species groups appear to have diverged no more than 23 000 years ago (Sota et al., 2013). The difference in life cycle length translates into a potential difference in fitness, because a 13-year life cycle has a 1.3 times greater intrinsic rate of population increase than a 17-year life cycle if all else is equal. However, extension of the juvenile period may affect fitness differently through its promotion of increased adult body size. Larger body size may be associated with greater fecundity in females (Karban, 1997) and increased mating success in males (Lloyd & Dybas, 1966; Karban, 1983). If adult body size does increase with an increase in life cycle length, 13-year cicadas may have smaller bodies and lower reproductive capacities than 17-year cicadas under the same habitat conditions, as is assumed in recent models of cicada evolution (Tanaka et al., 2009; Yoshimura et al., 2009). The correlation between body
body size [life cycle, y] species
Species group
(a)
2.906 2.907
Decula
2.906 2.903 Cassini
2.885 2.890
2.911 Decim 2.931
5
4
2 3 Age, mya
1
1271
[17] M. septendecula [13] M. tredecula [17] M. cassini
2.880
[13] M. tredecassini
2.928
[17] M. septendecim
2.918
[13] M. neotredecim
2.943
[13] M. tredecim
0
(b) 45°
40°
Cycle 17-y 13-y
35°
30°
95°
90°
85°
80°
75°
70°
Fig. 1 (a) Phylogeny and divergence times of Magicicada species (modified from Sota et al., 2013). Numerals at tree tips are mean body size (log10-transformed head width in 10-lm unit) of each species adjusted for sex and habitat annual mean temperature (AMT) (Table 2); numerals at tree nodes are reconstructed ancestral body sizes. (b) Sampling locations. Open and closed circles indicate 13- and 17-year cicadas, respectively.
size and life cycle length will hold if 13- and 17-year cicadas have the same annual nymphal growth rates under the same habitat conditions (Fig. 2 line 1). However, it is suggested that 17-year cicadas have a developmental arrest during the early juvenile period (White & Lloyd, 1975). If the additional 4 years of 17-year cicadas under the habitat conditions sufficient for 13-year life cycle are effectively a period of developmental arrest that adjusts the life cycle length to 17 years, 17-year cicadas would be expected to attain body sizes equivalent to those of 13-year cicadas (Fig. 2 line 2). Thus, the relatively recent putative life cycle shifts may have resulted in little difference in body size between the two life cycle lengths under the same habitat conditions, and the 13-year life cycle may have attained higher fitness than 17-year life cycle simply due to the shortened generation time. The relationship between life cycle length and body size has not been established by empirical studies except indirectly by studies of wing morphometric characters
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1272
T. KOYAMA ET AL.
Brood XIX (2011) and fixed them in 99% ethanol (Fig. 1b; Table S1). Note that specific brood numbers in roman numerals are assigned for cohorts of Magicicada that will emerge as adults at 13- or 17-year intervals (Marlatt, 1907). The specimens were stored at 30 to 20 °C. To estimate body size, we selected head width as measured by the distance between the outer edges of the eyes, because we determined that this measure could be taken reliably for every adult specimen. We measured the head width using digital vernier callipers for 4–65 samples from each location for a total of 4719 samples (Table S1). Explanatory variables
(Simon, 1983, 1992), which did not account for variation in relation to temperature/latitude. In this study, we investigated the adult body sizes of all seven Magicicada species throughout their geographic ranges to determine whether the species show clinal body size variation with geographic and climatic variables, and whether the body sizes of 13- and 17-year species within the same species group show any differences related to their life cycle differences as well as their history of divergence. Specifically, we tested whether the body sizes of 13- and 17-year cicadas at the same geographic location are similar to each other as expected if 17-year cicadas experience lower overall growth rates, as proposed by White & Lloyd (1975). We discuss the evolutionary processes and selective forces leading to the observed pattern of body size variation along a climatic gradient, between 13- and 17-year life cycles, and among species with different phylogenetic histories.
Potentially, a large number of geographic and climatic variables could be used in the analysis of insect body size variation across a geographic range. We initially tested latitude, longitude, altitude and multiple climatic variables included in WorldClim database (Hijmans et al., 2005; http://www.worldclim.org) as the variables to explain variation in body size. However, the WorldClim variables generally showed high correlations with one another and with latitude (Table S2). Such high correlations between explanatory variables caused the problem of multicollinearity and made the interpretation of statistical results difficult. Therefore, we decided to use geographic and climatic variables separately in different models. Further, we selected a single climatic variable, annual mean temperature (AMT) at each sampling location as the representative variable that reflects climatic condition. AMT data were obtained at a resolution of 30 s of latitude/longitude from WorldClim. We expect that AMT is related to forest productivity, on which the growth of cicada nymphs depends. Previous work by Cooley et al. (2013) showed that AMT was a strong predictor of 17-year cicada habitat suitability. We also have checked how AMT is related to ‘season length in day-degree unit’ measured by cumulative daily temperature above a developmental threshold temperature, which is often used as the explanatory climatic variable for clinal body size variation in insects. Because developmental threshold temperature for periodical cicadas is unknown, we obtained annual cumulative day-degrees above 0 °C using monthly mean temperatures for each geographic location. The season length thus obtained was highly correlated with AMT (n = 152 locations, r = 0.9926, P < 0.0001), and use of this variable instead of AMT yielded no substantial difference in statistical results.
Materials and methods
Statistical analyses
Fig. 2 Hypothetical growth curves for 13- and 17-year cicadas (body size is head width in logarithmic scale). Lines (1) represent constant growth curve shared by both 13- and 17-year cicadas; line (2) represents constant growth comparable to line 1 after a period of slow growth in early instars for 17-year cicadas.
Sampling and measurement of body size We collected 17-year cicadas of brood XIII (2007), XIV (2008), I (2012) and II (2013) and 13-year cicadas of
We conducted multiple regression analyses for the effects of sex, species group, species nested within species group, geographic variable (latitude, longitude, altitude) and AMT on body size (head width) using the
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1273
Body size variation in periodical cicadas
statistical package JMP version 10.0 (SAS Institute Inc., Cary, NC, USA). Head width data were normalized by log transformation, and mean values were calculated for each sex at each locality. In the multiple regression analysis, the dependent variable (mean head width) was weighted by sample size. Because latitude and AMT were strongly corrected with each other, we conducted analyses with geographic variables and AMT separately to avoid the problem of multicollinearity. In the model with geographic variables, all interaction terms with the geographic variables were initially incorporated and nonsignificant terms were removed in the final model. Following the multiple regression analysis with AMT, a Tukey HSD test was used to compare species’ mean body sizes (least square means). Also, we reconstructed body size at six ancestral nodes of the Magicicada phylogenetic tree from the species’ mean body sizes using ‘ace’ function of the package ‘ape’ (Paradis et al., 2004) in R version 2.15.2 (R Development Core Team, 2012). Comparison of nymphal instar compositions between 13- and 17-year cicadas To compare the developmental patterns during nymphal periods of 13- and 17-year cicadas, we analysed the instar composition data reported in White & Lloyd (1975) and Maier (1996). We tested the difference in instar composition between populations by G-test and the relationship between instar composition and habitat AMT by ordinal logistic regression. We used R version 2.15.2 for these statistical tests.
Results In the multiple regression models taking into account geographic or climatic factors, body size was negatively related to latitude and positively related to AMT (Table 1; Fig. 3). In the model with geographic vari-
Table 1 Multiple regression analysis for the effects of species group, species (nested in species group), sex, geographic variables and annual mean temperature (AMT) on body size (log10transformed head width). The effects of geographic variables and AMT were examined in separate models. Variable
d.f.
(a) Geographic variables Species group 2 Species 4 (species group) Sex 1 Latitude 1 Longitude 1 Altitude 1 Longitude*altitude 1 (b) AMT Species group 2 Species 4 (species group) Sex 1 AMT 1
F
P
Effect SE*
803.98 52.76
< 0.0001 < 0.0001
– –
605.10 60.20 0.26 1.33 10.95
< 0.0001 < 0.0001 0.6129 0.2495 0.0010
850.42 55.23
< 0.0001 < 0.0001
593.54 55.14
< 0.0001 < 0.0001
1.21E-02 2.04E-03 6.27E-05 5.79E-06 2.72E-06
4.93E-04 2.63E-04 1.24E-04 5.02E-06 8.22E-07
– – 1.22E-02 5.01E-04 2.25E-03 3.03E-04
*Effects for species group and species are not shown. See Table 2 for the result of multiple comparison of mean body sizes between species in the analysis on AMT (b).
ables (Table 1a), the effect of latitude was significant, as was the interaction effect of longitude and latitude. Having adjusted for these factors, sex, species group and species nested within species group showed consistently significant effects (Table 1). Females were larger than males in each species, demonstrating sexual size dimorphism (Table 2). The Decim group species were the largest, followed by species in the Decula group, and then those in the Cassini group (Table 2). In the Decim group, M. tredecim (13-year) was the largest, followed by M. septendecim (17-year) and then M. neotredecim (13-year). In the Decula group, body sizes did not differ between M. septendecula (17-year) and
Table 2 Mean head width (log10-transformed; in 10-lm unit) of each species and sex. Head width (least square mean SE) By species and sex† Species group
Species
Decim Decim Decim Decula Decula Cassini Cassini
Magicicada Magicicada Magicicada Magicicada Magicicada Magicicada Magicicada
tredecim septendecim neotredecim septendecula tredecula cassini tredecassini
Life cycle (years)
By species*
Male
13 17 13 17 13 17 13
2.943a 0.00137 2.928b 0.00131 2.918c 0.00145 2.907d 0.00181 2.906d 0.00233 2.890e 0.00126 2.880f 0.00123
2.931 2.919 2.909 2.889 2.888 2.877 2.866
Female
0.00172 0.00167 0.00185 0.00230 0.00331 0.00163 0.00161
2.954 2.937 2.926 2.926 2.922 2.902 2.893
0.00172 0.00169 0.00210 0.00266 0.00294 0.00162 0.00160
*Means adjusted for the effects of annual mean temperature (AMT) and sex following the multiple regression analysis given in Table 1b. Means followed by the same superscript letter do not differ significantly at P = 0.05 by Tukey HSD test. †Means by sex adjusted for the effect of AMT, showing that females are larger than males in each species. ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1274
T. KOYAMA ET AL.
M. tredecula (13-year). In the Cassini group, M. cassini (17-year) was significantly larger than M. tredecassini (13-year). In the analysis of nymphal instar composition data (Table 3), a 9-year-old 13-year population (Grand Tower) had a more advanced instar composition than a 9-year-old 17-year population at warmer habitat (Webber’s Falls) (G-test; G = 437.39, P < 2.2 9 10 16). However, instar composition did not differ significantly between 13- and 17-year populations 4 years before emergence (13-year-old Lewis Centre vs. 9-year-old Grand Tower populations: G-test, G = 4.14, P = 0.1264). (a) Decim male 2.98
(b)
M. septendecim M. neotredecim M. tredecim
2.93
2.88
2.88
2.83 7.0
11.0
15.0
19.0
(c) Cassini male
log10 Body size
2.98
2.98
2.93
2.88
2.88
15.0
19.0
(e) Decula male 2.98
15.0
19.0
M. cassini M. tredecassini
2.83 7.0
11.0
15.0
19.0
(f) Decula female
M. septendecula M. tredecula
2.98
2.93
M. septendecula M. tredecula
2.93
2.88
2.83 7.0
11.0
(d) Cassini female
M. cassini M. tredecassini
11.0
Our data reveal a complex pattern of sexually dimorphic body size variation related to climate, species (life cycle) and phylogeny. First, periodical cicadas showed a trend of increasing body size with decreasing
M. septendecim M. neotredecim M. tredecim
2.83 7.0
2.93
2.83 7.0
Discussion
Decim female
2.98
2.93
For the four 9th-year cohorts of 17-year cicadas, the effect of AMT on instar composition was significant (ordinal logistic regression: v2 = 185.54, P < 0.0001, b = 0.382), suggesting that nymphal development was slower in cooler habitats.
2.88
11.0
15.0
19.0
2.83 7.0
Annual mean temperature (°C)
11.0
15.0
19.0
Fig. 3 Relationship between body size (log10-transformed head width in 10lm unit) and annual mean temperature (AMT) by species for each sex in each species group. (a) Decim group, male; (b) Decim group, female; (c) Cassini group, male; (d) Cassini group, female; (e) Decula group, male; (f) Decula group, female. Bars indicate SE.
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Body size variation in periodical cicadas
1275
Table 3 Nymphal instar composition of 13- and 17-year cicada cohorts from six localities. Locality
Latitude (°N) Longitude (°W) Annual mean temperature (°C) Species Life cycle (year) Age at sampling (year) No. of nymphs 2nd instar 3rd instar 4th instar 5th instar Mean instar
Pilot Mound, IA*
Southington, CT†
Nashville, IN*
Webber’s Falls, OK*
Lewis Centre, OH*
Grand Tower, IL*
42.12 93.97 8.9
41.6 72.88 9.9
39.27 86.2 11.5
35.5 95.17 15.9
40.15 83.02 10.4
37.55 89.5 13.5
Magicicada cassini 17 9
Magicicada septendecim 17 9
Unknown
M. cassini
Unknown
Unknown
17 9
17 9
17 13
13 9
27 106 457 0 3.73
0 25 75 0 3.75
0 31 83 2 3.75
0 12 306 68 4.15
0 0 35 214 4.86
0 2 39 334 4.89
Source: *White & Lloyd (1975); †Maier (1996).
latitude (and increasing AMT), following the converse Bergmann’s rule (Table 1). This clinal pattern may indicate local adaptation to the length of growing season available for larval development; in southern (warmer) regions, annual time for development is longer, and large body size could be a result of selection for greater fecundity (Masaki, 1967; Roff, 1980; Karban, 1997). After removing climatic (AMT) effects, the anciently diverged 13-year M. tredecim showed the largest body size among the three Decim species. In the remaining six described species (excluding M. tredecim), 17-year species had similar or larger body sizes than their corresponding 13-year species after adjusting for the effect of AMT (Table 1; Fig. 3). This increase in body size is likely related to the 4-year extension of nymphal period of 17-year species compared to 13-year. However, the observed increase in body size (in log10 scale) was only 0.3–2.3% (Table 2). In contrast, based on the assumption that 17- and 13-year species grow at the same rate in log10 scale, 17-year cicadas are expected to be approximately 31% (4/13) larger than 17-year cicadas. Thus, 13-year cicadas must have faster growth rates on average than 17-year cicadas at the same AMT to achieve comparable body sizes. Both 13- and 17-year Magicicada species pass through five nymphal instars (White & Lloyd, 1975; Maier, 1996). Control of the nymphal period is unknown, but White & Lloyd (1975) postulated that 17-year cicadas spend longer nymphal periods in the early instars than do 13-year cicadas, implying that switching between 13- and 17-year cycles occurs by insertion/deletion of a 4-year period of developmental delay during the early juvenile period (Fig. 2). A full test of this hypothesis awaits age structure data throughout the entire juvenile periods for geographically close 13- and 17-year populations. For the
moment, our analysis of available data shows that nymphal instar composition 4 years before emergence in 13- and 17-year cicadas occurring at similar habitat temperatures did not differ whereas 9-year-old 13-year cicadas had a more advanced age structure than 9-year-old 17-year cicadas (Table 3). This suggests that the difference in development/growth rates between 13- and 17-year cicadas is attributable to the difference in early instars, specifically a slower development rate in early instars of 17-year species, as was postulated by White & Lloyd (1975). Thus, 17-year cicadas may experience a 4-year retardation of growth in early nymphal instars and thereafter grow as rapidly as 13-year cicadas to attain final body sizes similar to those of 13-year cicadas. In the Decim group, we found that the anciently diverged 13-year M. tredecim, unlike the recently diverged 13-year M. neotredecim, was larger than the 17-year M. septendecim (Table 1). This pattern may be the result of M. tredecim’s long-term adaptation to longer growing seasons given its generally southern distribution, while M. neotredecim individuals are small due to this species’ relatively recent divergence from M. septendecim (Marshall & Cooley, 2000; Simon et al., 2000; Sota et al., 2013) and the shorter growing seasons of the northern latitudes in which this species occurs. In conclusion, the body size differences between recently diverged 13- and 17-year species were negligible, while more anciently separated taxa in the Decim group showed notable differences, with M. tredecim larger than M. septendecim/neotredecim. Recent divergences between 13- and 17-year species in each species group (Sota et al., 2013) have been attributed to switching from the 17- to the 13-year cycle (e.g. Martin & Simon, 1988; Marshall & Cooley, 2000; Simon et al.,
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1276
T. KOYAMA ET AL.
2000; Cooley et al., 2001). The similarity in body size between recently diverged 13- and 17-year species in each species group suggests that along with the life cycle switch from 17- to 13-year cycles, there was a recovery of growth rates in early instars, which had been lowered in the 17-year cycle. Body size conservatism across a major life cycle shift appears to be unique to periodical cicadas. In insects, a drastic change in larval period length, such as from univoltine to bivoltine life cycles in crickets, is expected to result in a noticeable change in body size (Masaki, 1972; Roff, 1980; but see also Kivel€a et al., 2011 for counterexamples). Body size and associated dimensions may be related to other reproductive characters besides reproductive capacity such as male mating call frequency (Bennet-Clark & Young, 1994; but see Oberd€ orster & Grant, 2007); thus, a reduction in body size may reduce mating rate or likelihood of mating in addition to directly reducing reproductive capacity. Therefore, the body size conservatism in periodical cicadas during a life cycle switch may be adaptive for securing mating, although the resultant body size may not be optimal in terms of climatic adaptation.
Acknowledgments We thank H. Ikeda and D. Suzuki for sampling and W. Blanckenhorn for helpful comments on our manuscript. This study was supported by JSPS KAKENHI (no. 22255004, 22370010, 26257405 and 15H00420 to JY) and grants-in-aid for JSPS Fellows to HI and SK. CS and JRC received support from NSF DEB 09-55849.
References Alexander, R.D. & Moore, T.E. 1962. The evolutionary relationships of 17-year and 13-year cicadas, and three new species (Homoptera: Cicadidae, Magicicada). Univ. Mich. Mus. Zool. Misc. Pub. 121: 1–59. Bennet-Clark, H. & Young, D. 1994. The scaling of song frequency in cicadas. J. Exp. Biol. 191: 291–294. Blanckenhorn, W.U. & Demont, M. 2004. Bergmann and converse Bergmann latitudinal clines in arthropods: two ends of a continuum? Integr. Comp. Biol. 44: 413–424. Cooley, J.R., Simon, C., Marshall, D.C., Slon, K. & Ehrhardt, C. 2001. Allochronic speciation, secondary contact, and reproductive character displacement in periodical cicadas (Hemiptera: Magicicada spp.): genetic, morphological, and behavioural evidence. Mol. Ecol. 10: 661–671. Cooley, J.R., Kritsky, G., Edwards, M.J., Zyla, J.D., Marshall, D.C., Hill, K.B.R. et al. 2009. The distribution of periodical cicada brood X in 2004. Am. Entomol. 55: 106–112. Cooley, J.R., Marshall, D.C., Simon, C., Neckermann, M.L. & Bunker, G. 2013. At the limits: habitat stability modeling of northern 17-year periodical cicada extinctions (Hemiptera: Magicicada spp.). Glob. Ecol. Biogeogr. 22: 410–421. Gilchrist, G.W., Huey, R.B. & Serra, L. 2001. Rapid evolution of wing size clines in Drosophila subobscura. Genetica 112–113: 273–286.
Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G. & Jarvis, A. 2005. High resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25: 1965–1978. Huey, R.B., Gilchrist, G.W., Carlson, M.L., Berrigan, D. & Serra, L. 2000. Rapid evolution of a geographic cline in size in an introduced fly. Science 287: 308–309. Karban, R. 1983. Sexual selection, body size and sex-related mortality in the cidada Magiciada cassini. Am. Midl. Nat. 109: 324–330. Karban, R. 1997. Evolution of prolonged development: a life table analysis for periodical cicadas. Am. Nat. 150: 446–461. Kivel€a, S.M., V€alim€aki, P., Carrasco, D., M€aenp€a€a, M.I. & Oksanen, J. 2011. Latitudinal insect body size clines revisited: a critical evaluation of the saw-tooth model. J. Anim. Ecol. 80: 1184–1195. Lloyd, M. & Dybas, H.S. 1966. The periodical cicada problem. II. Evolution. Evolution 20: 466–505. Maier, C.T. 1996. Connecticut is awaiting return of the periodical cicada. Front. Plant Sci. 48: 4–6. Marlatt, C.L. 1907. The periodical cicada. US Dept. Agric. Bureau Entomol. Bull. 71: 1–183. Marshall, D.C. & Cooley, J.R. 2000. Reproductive character displacement and speciation in periodical cicadas, with description of a new species, 13-year Magicicada neotredecim. Evolution 54: 1313–1325. Martin, A.P. & Simon, C. 1988. Anomalous distribution of nuclear and mitochondrial DNA markers in periodical cicadas. Nature 336: 237–239. Masaki, S. 1967. Geographic variation and climatic adaptation in a field cricket (Orthoptera: Gryllidae). Evolution 21: 725– 741. Masaki, S. 1972. Climatic adaptation and photoperiodic response in the band-legged cricket. Evolution 26: 587–600. Oberd€ orster, U. & Grant, P.R. 2007. Acoustic adaptations of periodical cicadas (Hemiptera: Magicicada). Biol. J. Linn. Soc. 90: 15–24. Okuzaki, Y., Takami, Y. & Sota, T. 2010. Resource partitioning or reproductive isolation: the ecological role of body size differences among closely related species in sympatry. J. Anim. Ecol. 79: 383–392. Paradis, E., Claude, J. & Strimmer, K. 2004. APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20: 289–290. R Development Core Team 2012. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Roff, D. 1980. Optimizing development time in a seasonal environment: the ‘ups and downs’ of clinal variation. Oecologia 45: 202–208. Simon, C. 1983. Morphological differentiation in wing venation among broods of 13- and 17-year periodical cicadas. Evolution 37: 104–115. Simon, C. 1992. Discriminant analysis of year classes of periodical cicada based on wing morphometric data enhanced by molecular information. In: Ordinations in the Study of Morphology Evolution and Systematics of Insects: Applications and Quantitative Genetic Rationales (J.T. Sorensen & T.G. Footit, eds), pp. 309–322. Elsevier, Amsterdam, The Netherland. Simon, C., Tang, J., Dalwadi, S., Staley, G., Deniega, J. & Unnasch, T.R. 2000. Genetic evidence for assortative mating between 13-year cicadas and sympatric “17-year cicadas
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Body size variation in periodical cicadas
with 13-year life cycles” provides evidence for allochronic speciation. Evolution 54: 1326–1336. Sota, T., Takami, Y., Kubota, K. & Ishikawa, R. 2000a. Geographic variation in the body size of some Japanese Leptocarabus species (Coleoptera, Carabidae): the “toppleddomino pattern” in species along a geographic cline. Entomol. Sci. 3: 309–320. Sota, T., Takami, Y., Kubota, K., Ujiie, M. & Ishikawa, R. 2000b. Interspecific body size differentiation in species assemblage of the carabid subgenus Ohomopterus in Japan. Popul. Ecol. 42: 279–291. Sota, T., Yamamoto, S., Cooley, J.R., Hill, K.B.R., Simon, C. & Yoshimura, J. 2013. Independent divergence of 13- and 17-y life cycles among three periodical cicada lineages. Proc. Natl. Acad. Sci. USA 110: 6919–6924. Tanaka, Y., Yoshimura, J., Simon, C., Cooley, J.R. & Tainaka, K. 2009. Allee effect in the selection for prime-numbered cycles in periodical cicadas. Proc. Natl. Acad. Sci. USA 106: 8975–8979. White, J. & Lloyd, M. 1975. Growth-rates of 17-year and 13year periodical cicadas. Am. Midl. Nat. 94: 127–143.
1277
Yoshimura, J., Hayashi, T., Tanaka, Y., Tainaka, K. & Simon, C. 2009. Selection for prime-number intervals in a numerical model of periodical cicada evolution. Evolution 63: 288– 294.
Supporting information Additional Supporting Information may be found in the online version of this article: Table S1 Sampling locations of each species. Table S2 Correlation coefficients between latitude, annual mean temperature and other climate parameters for sampling locations in this study. Data deposited at Dryad: doi:10.5061/dryad.mn200 Received 8 February 2015; revised 24 February 2015; accepted 20 April 2015
ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY. J. EVOL. BIOL. 28 (2015) 1270–1277 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2015 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
