bs_bs_banner

Biological Journal of the Linnean Society, 2013, , . With 7 gures

Variance in developmental event timing is greatest at low biological levels: implications for heterochrony
OLIVER TILLS*, SIMON D. RUNDLE and JOHN I. SPICER
Marine Biology and Ecology Research Centre, School of Marine Science and Engineering, Plymouth University, Drake Circus, Plymouth, Devon PL4 8AA, UK
Received 21 June 2013; accepted for publication 3 July 2013

Intraspecic variation in developmental event timing is common and may be the raw material from which heterochronies (altered timing of developmental events between ancestors and descendants) arise. However, our understanding of how variance in intraspecic developmental event timing is distributed across different hierarchical, biological levels is poor, despite recent evidence suggesting a genetic basis for such inter-individual differences. In the present study, we used high (temporal and spatial) resolution bio-imaging of the entire embryonic development of the pond snail, Radix balthica, to investigate how variance was partitioned between the biological levels of interpopulation, inter-egg mass and individual, with respect to the relative timing of four key developmental events (four-cell division, a discrete heart beat, capsule rupture, and hatching), as well as egg volume (an index of maternal investment), hatchling size and shape (to compare embryonic with post-hatch variance). We found that the timing of all but one (four cell division) of the developmental events, together with all measures of hatchling size and shape, had most variance partitioned at the individual level; variance at the interpopulation level was surprisingly low despite sampling populations from across the geographical range of R. balthica. These patterns highlight the importance of studying the distribution of variance in developmental event timing with the aim of understanding how it might drive organismal ecology and evolution. 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, , .

ADDITIONAL KEYWORDS: inter-individual variation Mollusca pond snail Radix balthica variance
partitioning.

INTRODUCTION
Variation is the raw material required for natural selection to operate (Darwin, 1859) and understanding the nature of variation in physiological, morphological, developmental, and behavioural traits is key to understanding ecology and evolution (Spicer & Gaston, 1999; Arthur, 2011). To investigate the biological signicance of variation, the pattern that this variation takes in nature should be documented (Spicer & Gaston, 1999; Chown & Gaston, 2008). However, given that the process of quantifying variance in a particular trait at different biological levels is simple, compared to investigating the processes underlying variation, it is an approach that is underutilized in evolutionary biology. Hence, although

*Corresponding author. E-mail: oliver.tills@plymouth.ac.uk

there are studies documenting variance partitioning (Heatwole, Mercado & Oritz, 1965; Chown, Le Lagadec & Scholtz, 1999; Bagatto, Crossley & Burggren, 2000), they are not common, particularly for lower biological levels (e.g. population, clutch, inter-individual). Those studies that have investigated the partitioning of variance in a trait have provided novel insight into trait ecology and evolution. For example, Chown et al. (1999) investigated variance partitioning at the levels of genus, species, population, and individual for traits related to desiccation resistance in African keratin beetles. Once the effects of body size had been accounted for, variance in physiological traits such as maximum tolerable water loss and lipid and water content was highest at the individual level. Bagatto et al. (2000) also documented the distribution of variance in physiology, this time in neo-natal nine-banded armadillos. This species produces monozygous

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

O. TILLS ET AL. inter- and intraspecic levels (Smirthwaite et al., 2007; Tills, Spicer & Rundle, 2010; Tills et al., 2011), is partitioned at the levels of interpopulation, interegg mass, and individual. Given the potential inuence of maternal investment on development (Ho & Burggren, 2010), we investigated variance partitioning in egg volume and the effect of including this trait, as a covariate, on the partitioning of variance in developmental event timing. Finally, we compared variance partitioning in hatchling size and shape with that for embryonic traits. Our overall aim was to understand how variation in these early developmental traits was partitioned at different biological levels, which represents a key step in understanding what intraspecic variation in developmental event timing means for heterochrony.

quadruplets and therefore, within a litter, there is zero genetic variability. Bagatto et al. (2000) revealed that between-litter variability was always greater than that present within individual litters, suggesting a genetic component underlies this variation in early-life history. Although the above studies show that the investigation of variance partitioning can provide important insights for evolutionary biology, important questions remain. For example, we know little about variation at different biological levels during development; indeed, variation during development has often been viewed as noise that hampers the investigation of the genetic control of developmental processes (Spicer & Gaston, 1999; Spicer, Rundle & Tills, 2011). However, an emerging body of evidence is revealing signicant intraspecic variation during development, including variation in the timing of developmental events at the individual level (Gomez-Mestre et al., 2008, 2010; de Jong et al., 2009; Mourabit et al., 2010; Pan & Burggren, 2010; Tills et al., 2011; Tills, Rundle & Spicer, 2013). Altered timing of developmental events between ancestors and descendants, termed heterochrony, is a well-documented macroevolutionary pattern (McNamara, 1995; Spicer, 2006; Spicer et al., 2011; Bhuller et al., 2012), which was proposed by Gould (1977) to be the main mechanism of evolutionary innovation. Most studies of heterochrony have concentrated on documenting differences in developmental event timing between species, with an implicit assumption that any within-species variation is minor, and probably not relevant to addressing questions relating to this macroevolutionary pattern. However, evidence for intraspecic differences in developmental event timing is accumulating (Cubbage & Mabee, 1996; Mabee & Trendler, 1996; Mabee, Olmstead & Cubbage, 2000; Schmidt & Starck, 2004; Sheil & Greenbaum, 2005; de Jong et al., 2009; Kawajiri, Kokita & Yamahira, 2009), in addition to the ndings of a study by Tills et al. (2011, 2013) highlighting that such inter-individual variation could have a genetic basis. Given that intraspecic variation in developmental event timing may be biologically important, it is timely to begin addressing how this variation is partitioned at different biological levels to begin to appreciate the implications of such intraspecic variation for our understanding of heterochrony. A limitation in any study of developmental event timing is the temporal frequency with which an organism is observed. In the present study, we use high (temporal and spatial) resolution bio-imaging of the entire embryonic development of the snail Radix balthica to investigate, in unprecedented detail, how variance in the timing of ve key embryonic developmental events, which exhibit variation at both the

MATERIAL AND METHODS EMBRYO CULTURE

Second-generation stock populations of R. balthica from each of four European populations spanning this species latitudinal range (north Sweden: 66.428, 19.683; south-west England: 50.439, 3.690; Bavaria, Germany: 50.007, 9.156; south France: 44.053, 4.784) (Pfenninger et al., 2011) were maintained at 15 C under a 12 : 12 h light/dark cycle. Snails were cultured in aquaria (15 L; stocking density = 15) containing articial pond water (Rundle et al., 2004) and fed lettuce and spinach ad libitium. Egg masses were harvested from aquaria and examined under low power magnication (10) and, if eggs had not developed past the rst cell division, they were used in the experiments described below.

BIO-IMAGING
Eggs were dissected from egg masses under low power magnication (10) and six eggs were selected haphazardly from along the length of the egg mass (France: seven egg masses; Sweden: four egg masses; Germany: three egg masses; England: six egg masses) and cultured at 20 C in a 384-well microtitre plate (well volume = 84 L). The entire development of embryos was recorded using a custom, automated imaging system designed in our laboratory for imaging aquatic animals (Fig. 1). This system comprises a four megapixel shutterless monochrome camera (Pike 421 B; Allied Vision Technologies) connected to a zooming lens system (Zoom 70 XL; Optem), inverted beneath a motorized XY stage (Optiscan XY stage; Prior Scientic). Transmitted, dark-eld, cold lighting was provided by light emitting diodes. The camera and XY motorized stage were controlled using the open-source software MICROMANAGER, version 1.3 (Edelstein et al.,

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

PARTITIONING OF VARIANCE IN DEVELOPMENTAL EVENT TIMING

Figure 2. Mean SE absolute and relative (to the period between the two-cell divison and rst radula function) timing of the developmental events: four-cell division, identiable heart beat, capsule rupture, and hatching.

Figure 1. Drawing of the bio-imaging system used in the present study (designed and built in our laboratory) for imaging of aquatic embryos with high spatial and temporal resolution.

2010) run on a Mac Pro using OS X. The rst 4 days of development were captured by recording a single image of each embryo every 30 min. After day 4, image sequences were acquired at 7.5 frames s1 for 20 s, every 60 min. Articial pond water in each well of the multiwell plate was replaced daily.

IMAGE

ANALYSIS

Following image acquisition, image sequences were analysed by manual observation using IMAGE J (Abramoff et al., 2004). The time of onset of six embryonic developmental events (two-cell division, four-cell division, a discrete heart beat, radula ontogeny, egg capsule rupture, and hatching) was recorded for each embryo. The events used, have all been shown to vary in the timing of their appearance both within, (Tills et al., 2010; 2011) and between (Smirthwaite et al., 2007) species. The time period between two-cell division and radula ontogeny was used to standardize the timing of the remaining four developmental events

(four-cell division, a discrete heart beat, egg capsule rupture, and hatching) to provide a measure of relative event timing and therefore control for differences in the overall rate of development between embryos (Fig. 2). As a result of this standardization, two-cell division and radula ontogeny had no variance in relative timing and so were excluded from subsequent analyses. Radula ontogeny was chosen as the second developmental event with which to standardize developmental timing because it exhibits far less variation in timing than other late developmental events such as hatching (O. Tills, pers. observ.) and so its use produces a more consistent measure of relative developmental event timing. The resultant values for the relative timing of four-cell division, a discrete heart beat, egg capsule rupture, and hatching were used in subsequent statistical analyses (Fig. 3). Egg volume was calculated using the formula of Taylor (1973): 1/6lw2, where l is egg length and w is egg width measured from images obtained at the beginning of development. The shell parameters length, width, aperture length, aperture width, and aperture height were measured immediately after hatching. Aspect ratio (length : width), aperture ratio (aperture length : aperture width) and the ratio of length : aperture height, were calculated as measures of shell shape (Tills et al., 2010; Table S1).

STATISTICAL

ANALYSIS

A nested analysis of variance, using the Satterthwaite approximation (Sokal & Rohlf, 1995) for unequal

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

O. TILLS ET AL.

Figure 3. Micrograph showing Radix balthica embryos at the onset of the developmental events: two-cell division, four-cell division, identiable heart beat, radula, and capsule rupture.

Figure 4. Variance in egg volume (%) and the relative time of onset of developmental events partitioned at the population, egg mass and individual level in the pond snail, Radix balthica. Percentages were calculated both with and without egg volume included as a covariate.

sample sizes, was performed to partition variance in the relative timing of the four developmental events (four-cell division, a discrete heart beat, capsule rupture, hatching), egg volume, and hatchling shell size and shape parameters between the levels of interpopulation, inter-egg mass and individual (including the error term).

DEVELOPMENTAL

EVENT TIMING

RESULTS EGG VOLUME

Egg volume was signicantly different between egg masses (F17, 63 = 16.94, P 0.001) but not populations (F3, 63 = 1.69, P = 0.207). Most variation in egg volume was partitioned at the inter-egg mass level (72%) and the least at the interpopulation level (11%) (Fig. 4). Egg volume was a signicant covariate when included in the analysis of variance (ANOVA) model and therefore variance was calculated both with (analysis of covariance) and without egg volume (ANOVA) as a covariate to understand the partitioning of variance both with and without controlling for its effects.

Variance in the relative timing of all four measured developmental events was lowest at the interpopulation level, both with and without (Fig. 4) egg volume factored into the analyses. With egg volume as a covariate in the ANOVA model, most variation in the timing of developmental events was partitioned at the individual level; however, ANOVA without egg volume as a covariate revealed the timing of four-cell division had more variance partitioned at the inter-egg mass (39%) than at the individual level (34%). The partitioning of variance at the interpopulation level decreased as embryonic development progressed (i.e. four-cell division > heart beat > capsule rupture > radula > hatching), whereas the partitioning of variance at the individual level increased (Fig. 4). The relative timing of four-cell division, the earliest developmental event analyzed, was signicantly different between both populations and egg masses, whereas timing of an identiable heart beat, capsule rupture, and hatching was signicantly different between egg masses but not populations (Tables 1, 2).

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

PARTITIONING OF VARIANCE IN DEVELOPMENTAL EVENT TIMING

Table 1. Results of analysis of variance testing for differences in the relative time of onset of key embryonic developmental events between populations and egg masses in the pond snail Radix balthica Factor Four-cell division Population Egg mass Error A discrete heart beat Population Egg mass Error Capsule rupture Population Egg mass Error Hatching Population Egg mass Error d.f. SS MS F P

3 17 63 3 17 63 3 17 63 3 17 63

0.0001802 0.0002626 0.0001784 0.038403 0.097698 0.098787 0.40546 1.49737 1.71081 0.37815 1.43682 1.69404

0.0000601 0.0000154 0.0000028 0.012801 0.005747 0.001568 0.13515 0.08808 0.02716 0.12605 0.08452 0.02689

3.88 5.46

0.028 0.001

* **

2.23 3.67

0.122 0.001

NS **

1.53 3.24

0.242 0.001

NS **

1.49 3.14

0.253 0.001

NS **

NS, P > 0.05; *P < 0.05; **P 0.001.

Table 2. Results of analysis of covariance testing for differences in the absolute time of onset of key embryonic developmental events between populations and egg masses in the pond snail Radix balthica, with egg volume as a covariate Factor Four-cell division Egg volume Population Egg mass Error A discrete heart beat Egg volume Population Egg mass Error Capsule rupture Egg volume Population Egg mass Error Hatching Egg volume Population Egg mass Error d.f. SS MS F P

1 3 17 62 1 3 17 62 1 3 17 62 1 3 17 62

0.0000002 0.0000870 0.00001698 0.0001781 0.001846 0.038546 0.094898 0.096941 0.00313 0.22627 1.23417 1.70767 0.00223 0.20813 1.18079 1.69181

0.0000002 0.0000290 0.0000100 0.0000029 0.001846 0.012849 0.005582 0.001564 0.00313 0.07542 0.07260 0.02754 0.00223 0.06938 0.06946 0.02729

0.08 3.24 3.48

0.7785 0.046 0.001

NS * **

1.18 2.57 3.57

0.281 0.085 0.001

NS NS **

0.11 1.14 2.64

0.737 0.357 0.003

NS NS *

0.08 1.10 2.55

0.776 0.375 0.004

NS NS *

NS, P > 0.05; *P < 0.05; **P 0.001.

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

O. TILLS ET AL.

Figure 5. Variance (%) in hatchling shell parameters (L, length; W, width; AL, aperture length; AW, aperture width; AH, aperture height; L : W, aspect ratio; AR, aperture ratio; AH : L, ratio of aperture height to length) partitioned at the levels of population, egg mass and individual level in the pond snail, Radix balthica. Percentages are calculated both with and without egg volume included as a covariate.

HATCHLING

SIZE AND SHAPE

Most variation in size and shape parameters was partitioned at the individual level, both with (68%100%; Fig. 5) and without (71%98%; Fig. 5) egg volume as a covariate. There were no signicant differences between either populations or egg masses in any of the size and shape parameters measured.

DISCUSSION
The present study presents some of the rst evidence regarding how intraspecic variation in developmental event timing (the potential raw material from which heterochronies form) is distributed at different hierarchical levels across a species range. Variance partitioning in the timing of all embryonic developmental events after the four-cell division stage was greater at the individual level (5261%) than at the levels of either inter-egg mass (3436%) or interpopulation (125%) (Fig. 4). Egg volume when included as a covariate in the analysis decreased variance partitioning in developmental traits at both interpopulation and inter-egg mass levels, whereas hatchling size and shape parameters had most variance partitioned at the individual level, regardless of whether egg volume was included as a covariate (Figs 4, 5). An assumption in many studies of heterochrony is that intraspecic variation in developmental event timing is both small and inappropriate for addressing questions relating to this macro-evolutionary pattern (Spicer et al., 2011). In the present study, we show not only the prevalence of intraspecic variation in developmental event timing, but also that, with greater variation between individuals than between populations, variance partitioning was not as might have been predicted. The predominance of individual level variation in the developmental events studied could either have a genetic basis or be attributable to variation in environmental conditions during

development (Spicer & Gaston, 1999). Tills et al. (2011) showed that individual variation in the timing of developmental events within a population of R. balthica appeared to have a genetic basis; hence, some portion of the variance that we observed at the individual level might result from genetic differences. This previous study also showed that embryos from within a single egg mass were no more closely related to each other than to embryos from other egg masses within their population. Radix balthica is a mixed mating simultaneous hermaphrodite (Haun et al., 2012) and therefore a single egg mass can contain eggs produced using sperm from more than one father (Jarne & Delay, 1990), perhaps explaining the low levels of genetic differentiation between egg masses. If some portion of the variance uncovered in the present study has a genetic component, this population structure may help explain why there was not more partitioning of variance between egg masses. Radix balthica embryos exhibit plasticity in the timing of developmental events in response to both biotic (Rundle et al., 2011) and abiotic (Tills et al., 2010) stressors. No changes were intentionally made to the environmental conditions in which either parents or embryos were cultured; however, differences in egg volume, a surrogate for maternal provisioning, may have contributed to some of the variance recorded. When egg volume was included as a covariate in the ANOVA model, inter-egg mass variance in developmental event timing decreased, suggesting that egg volume was driving some of the variance partitioning at the egg mass level. For example, the distribution of variance in the relative timing of four-cell division was strongly inuenced by egg volume: without egg volume as a covariate, fourcell division had most variance at the egg mass level (39%); when egg volume was included as a covariate, however, most variance was partitioned at the individual level (48%). In the present study, embryos were

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

PARTITIONING OF VARIANCE IN DEVELOPMENTAL EVENT TIMING harvested from F2 stock populations in which variation in maternal age might exist, and this could have contributed both to variation in egg volume between egg masses and subsequent variation in embryonic developmental traits. Egg size has been shown to be affected by factors including: maternal age [both positive (George, 1994; Chester, 1996; Bingham, Giles & Jaeckle, 2004) and negative (Qian & Chia, 1992; Ito, 1997) relationships], nutrition (Bertram & Strathmann, 1998) and exposure to toxicants (Lardies, Medina & Correa, 2008). The potential inuence of egg volume was further investigated using regression analyses, which revealed that it was a good predictor of the relative timing of four-cell division and rst heart beat, as well as the size of hatchlings (Figs 6, 7). The positive relationship between hatchling size and egg volume appears to have been driven by both a longer overall embryonic period in larger eggs and also a relatively longer period of time between the radula becoming functional and hatching (O. Tills, pers. observ.). Most variance in hatchling size and shape parameters was partitioned at the individual level (length: 73%; width: 74%), whereas egg volume had most variance partitioned at the inter-egg mass level and was also signicantly different between egg masses. However, despite differences in egg volume between egg masses, there were no differences between egg masses

in the size of hatchlings. The cause of this apparent disparity between egg volume and hatchling size is unclear, although it perhaps results from individual variation in aspects of development that subsequently cause different energetic requirements among embryos from the same egg mass or, alternatively, perhaps variation in the composition of maternal provisioning (rather than simply the quantity) results in variation in the size of hatchlings from within an egg mass. Baur (1994) showed that eggs of the land snail, Arianta arbustorum, had less variable nutrient content than egg volume and that both nutrient content (nitrogen and carbon) and egg volume were more variable between, than within, egg masses. Baur (1994) also reported that only 61% of egg masses contained eggs with levels of nitrogen correlated with their volume. In the present study, although egg volume can be used as an indicator of maternal investment, it is unlikely to provide a complete measure. Egg size has also been shown to be heritable in the serpulid polychaete, Hydroides elegans, in response to articial selection, using a half-sib breeding analysis (Miles, Hadeld & Wayne, 2007). In the present study, egg volume affected both the size of hatchlings and the timing of several developmental events and, therefore, if some component of egg volume has a

Figure 6. Regression analysis of the relationship between egg volume and (A) the relative timing of four-cell division (R2 = 33.6, F1,82 = 42.94, P 0.001: y = 0.16 0.14x) and (B) appearance of a discrete heart beat (R2 = 19.0, F2,81 = 10.7, P 0.001: y = 0.78 0.13x) in the pond snail, Radix balthica. Models are those of best t.

Figure 7. Regression analysis of the effect of egg volume on the (A) length (R2 = 14.9, F1,54 = 5.9, P = 0.005: y = 0.6 + 1.28x 0.78x2) and (B) width (R2 = 11.0, F1,55 = 7.9, P = 0.007: y = 0.57 + 0.27x) of hatchlings, in the pond snail, Radix balthica. Models are those of best t.
2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

O. TILLS ET AL. ation in the traits that we did use being largely partitioned at the individual level raises important questions regarding variance partitioning in developmental timing and what this may mean for the link between intraspecic variance in developmental timing and heterochrony. Intraspecic variation in developmental event timing is a possible source of heterochrony. Hence, understanding the distribution of this intraspecic variance (both hierarchically and geographically) is of paramount importance for understanding the ecology and evolution of event timing and, ultimately, understanding how heterochronies occur. Individual variation in embryonic developmental timing appears not only to be prevalent in this species and to have a genetic basis but is still the predominant level of variation, even when studying populations from across the species geographical range. The implications of the variance partitioning reported in the present study about the relationship between developmental event timing at the inter- and intraspecic levels are unclear, although they highlight the need to address how variation in developmental event timing between these two levels is related.

genetic component, this may have important developmental and ecological consequences. The least variation in all of the developmental events measured in the present study was at the level of interpopulation despite the use of populations from the extremes of the geographical range of R. balthica (northern Sweden to southern France) (Pfenninger et al., 2011). This is surprising because, if a genetic basis for developmental differences can arise within a single population (Tills et al., 2011), it might be expected that such differences would also be evident between populations and therefore that a signicant portion of variation would be distributed at this level. However, Pfenninger et al. (2011), again using populations from across the range of R. balthica, demonstrated that there was no signicant relationship between interpopulation geographical distance and their genetic distance. A key dispersal mechanism for R. balthica is assumed to be transport via waterfowl and therefore the magnitude of dispersal can be fairly independent of distance. Pfenninger et al. (2011) attributed the absence of a relationship between geographical and genetic distance to the genetic variability that results from bird-mediated transport (Haun et al., 2012). Although the ndings of Pfenninger et al. (2011) help to explain why variation between populations might be lower than otherwise expected, it is still surprising given the results of Tills et al. (2011) demonstrating developmental timing differences related to genetic distance within a single population (from a 10 m stretch of shore), that there was not more partitioning of variance between populations. Variance between populations was greatest early in development (i.e. in the timing of four-cell division) and decreased as development progressed. Differences between populations in the timing of early development, which are subsequently replaced by variation mainly at the individual level in later development, are interesting because they suggest some mechanism by which individual variation increases as development progresses. The study by De Jong et al. (2009) perhaps best highlights the potential for considerable individual variation in developmental event timing, using an investigation of 82 developmental characters within Haplochromis piceatus, a species of Lake Victoria cichlid, using ontogenetic sequence analysis (Colbert & Rowe, 2008). Data from 261 embryos revealed 26 880 equally parsimonious developmental sequences, which is an incredible complexity of intraspecic developmental variation. In the present study, we employed relatively few embryonic developmental events. Although studying more traits may have revealed different patterns of variance distribution between groups of traits, the prevalence of vari-

ACKNOWLEDGEMENTS
This study was supported by a studentship to O.T. from Plymouth University. The development of the bio-imaging technology was supported by Plymouth Universitys Marine Institute, Research and Innovation Department and the Faculty of Science and Technology. We thank three reviewers for their helpful comments.

REFERENCES
Abramoff MD, Magalhes PJ, Paulo J, Ram SJ. 2004. Image processing with ImageJ. Biophotonics International 11: 3642. Arthur W. 2011. Searching for evo-devos Holy Grail: the nature of developmental variation. Evolution and Development 13: 405407. Bagatto B, Crossley DA, Burggren WW. 2000. Physiological variability in neonatal armadillo quadruplets: withinand between-litter differences. Journal of Experimental Biology 203: 17331740. Baur A. 1994. Within- and between-clutch variation in egg size and nutrient content in the land snail Arianta arbustorum. Functional Ecology 8: 581586. Bertram DF, Strathmann RR. 1998. Effects of maternal and larval nutrition on growth and form of planktotrophic larvae. Ecology 79: 315327. Bhullar BS, Marugn-Lobn J, Racimo F, Bever GS, Rowe TB, Norell MA, Abzhanov A. 2012. Birds have paedomorphic dinosaur skulls. Nature 487: 223226.

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

PARTITIONING OF VARIANCE IN DEVELOPMENTAL EVENT TIMING

Bingham BL, Giles K, Jaeckle W. 2004. Variability in broods of the seastar Lepasterias aequalis. Canadian Journal of Zoology 82: 457463. Chester CM. 1996. The effect of adult nutrition on the reproduction and development of the estuarine nudibranch, Tenellia adspersa (Nordmann, 1845). Journal of Experimental Marine Biology and Ecology 198: 113130. Chown SL, Gaston KJ. 2008. Macrophysiology for a changing world. Proceedings of the Royal Society of London Series B, Biological Sciences 275: 14691478. Chown SL, Le Lagadec MD, Scholtz CH. 1999. Partitioning variance in a physiological trait: dessication resistance in keratin beetles (Coleoptera, Trogidae). Functional Ecology 13: 838844. Colbert MW, Rowe T. 2008. Ontogenetic sequence analysis: using parsimony to characterize developmental sequences and sequence polymorphism. Journal of Experimental Zoology 310B: 398416. Cubbage CC, Mabee PM. 1996. Development of the cranium and paired ns in the zebrash Danio rerio (Ostariophysi, Cyprinidae). Journal of Morphology 229: 121160. Darwin C. 1859. On the origin of species, or the preservation of favoured races in the struggle for life. London: John Murray. De Jong IML, Colbert MW, Witte F, Richardson MK. 2009. Polymorphism in developmental timing: intraspecic heterochrony in a Lake Victoria cichlid. Evolution and Development 11: 625635. Edelstein A, Amodaj N, Hoover K, Vale R, Stuurman N. 2010. Computer control of microscopes using Manager. Current Protocols in Molecular Biology 92: 14.20.114.20.17. George SB. 1994. Population differences in maternal size and offspring quality for Lepasterias epichlora (Brandt) (Echinodermata: Asteroidea). Journal of Experimental Marine Biology and Ecology 175: 121131. Gomez-Mestre I, Saccoccio VL, Iljima T, Collines EM, Rosenthal GG, Warkentin KM. 2010. The shape of things to come: linking developmental plasticity to postmetamorphic morphology in anurans. Journal of Evolutionary Biology 23: 13641373. Gomez-Mestre I, Wines JJ, Warkentin KM. 2008. Evolution of adaptive plasticity: risk sensitive hatching in neotropical leaf breeding treefrogs. Ecological Monographs 78: 205 224. Gould SJ. 1977. Ontogeny and phylogeny. Cambridge, MA: Harvard University Press. Haun T, Salinger M, Pachzellt A, Pfenninger M. 2012. On the processes shaping small-scale population structure in Radix balthica (Linnaeus 1758). Malacologia 55: 219233. Heatwole H, Mercado N, Oritz E. 1965. Comparison of critical thermal maxima of two species of Puerto Rican frogs of the genus Eleutherodactylus. Physiological Zoology 38: 18. Ho DH, Burggren WW. 2010. Epigenetics and transgenerational transfer: a physiological perspective. Journal of Experimental Biology 213: 316. Ito K. 1997. Egg-size and number variations related to maternal size and age, and the relationship between egg size and larval characteristics in an annual marine gastropod, Haloa

japonica (Opisthobranchia; Cephalaspidea). Marine Ecology Progress Series 152: 187195. Jarne P, Delay B. 1990. Inbreeding depression and selffertilization in Lymnaea peregra (Gastropoda: Pulmonata). Heredity 64: 169175. Kawajiri M, Kokita T, Yamahira K. 2009. Heterochronic differences in n development between latitudinal populations of the medaka Oryzias latipes (Actinopterygii: Adrianichthyidae). Biological Journal of the Linnean Society 97: 571580. Lardies MA, Medina MH, Correa JA. 2008. Intraspecic biogeographical pattern breakage in the snapping shrimp Betaeus emarginatus caused by coastal copper mine tailings. Marine Ecology Progress Series 358: 203210. Mabee PM, Olmstead KL, Cubbage CC. 2000. An experimental study of intraspecic variation, developmental timing and heterochrony in shes. Evolution 54: 20912106. Mabee PM, Trendler TA. 1996. Development of the cranium and paired ns in Betta splendens (Teleostei: Percomorpha): intraspecic variation and interspecic comparisons. Journal of Morphology 227: 249287. McNamara KJ. 1995. Evolutionary change and heterochrony. Chichester: Wiley. Miles CM, Hadeld MG, Wayne HL. 2007. Heritability for egg size in the serpulid polychaete Hydroides elegans. Marine Ecology Progress Series 340: 155162. Mourabit S, Rundle SD, Spicer JI, Sloman KA. 2010. Alarm substance from adult zebrash alters early embryonic development in offspring. Biology Letters 6: 525 528. Pan TCF, Burggren WW. 2010. Onset and early development of hypoxic ventilator responses and branchial neuroepithelial cells in Xenopus laevis. Comparative Biochemistry and Physiology 157: 382391. Pfenninger M, Salinger M, Haun T, Feldmeyer B. 2011. Factors and processes shaping the population structure and distribution of genetic variation across the species range of the freshwater snail Radix balthica (Pulmonata, Basommatophora). BMC Biotechnology 11: 135. Qian PY, Chia FS. 1992. Effect of aging on reproduction in a marine polychaete Capitella sp. Journal of Experimental Marine Biology and Ecology 156: 2338. Rundle SD, Smirthwaite JJ, Colbert MW, Spicer JI. 2011. Predator cues alter the timing of developmental events in gastropod embryos. Biology Letters 7: 285287. Rundle SD, Spicer JI, Coleman RA, Vosper J, Soane J. 2004. Environmental calcium modies induced defenses in a freshwater gastropod. Proceedings of the Royal Society London Series B, Biological Sciences 271: 6770. Schmidt K, Starck JM. 2004. Developmental variability during early embryonic development of zebra sh, Danio rerio. Journal of Experimental Zoology 302B: 446457. Sheil CA, Greenbaum E. 2005. Reconsideration of skeletal development of Chelydra serpentina (Reptilia: Testudinata: Chelydridae): evidence for intraspecic variation. Journal of Zoology 265: 235267. Smirthwaite JJ, Rundle SD, Bininda-Emonds ORP, Spicer JI. 2007. An integrative approach identies

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,

10

O. TILLS ET AL.
Taylor HH. 1973. The ionic properties of the capsular uid bathing embryos of Lymnaea stagnalis and Biomphalaria sudanica (Mollusca: Pulmonata). Journal of Experimental Biology 59: 543564. Tills O, Rundle SD, Salinger M, Haun T, Pfenninger M, Spicer JI. 2011. A genetic basis for intraspecic differences in developmental timing? Evolution and Development 13: 542548. Tills O, Rundle SD, Spicer JI. 2013. Parentoffspring similarity in the timing of developmental events: an origin of heterochrony? Proceedings of the Royal Society: Biological Sciences 280: 1769. Tills O, Spicer JI, Rundle SD. 2010. Salinity-induced heterokairy in an upper-estuarine population of the snail Radix balthica (Mollusca: Pulmonata). Aquatic Biology 9: 95105.

developmental sequence heterochronies in freshwater basommatophoran snails. Evolution and Development 9: 122130. Sokal RR, Rohlf FJ. 1995. Biometry. The principles and practice of statistics in biological research, 3rd edn. New York, NY: WH Freeman. Spicer JI. 2006. A physiological approach to heterochrony. In: Warburton SJ, Burggren WW, Pelster B, Reiber CL, Spicer J, eds. Comparative developmental physiology: contributions, tools, and trends. USA: Oxford University Press. Spicer JI, Gaston KJ. 1999. Physiological diversity and its ecological implications. Oxford: Blackwell Science. Spicer JI, Rundle SD, Tills O. 2011. Studying the altered timing of physiological events during development: its about time . . . or is it? Respiratory Physiology and Neurobiology 178: 312.

SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article at the publishers web-site: Table S1. Embryonic developmental event timing (h), hatchling size and shape (mm) and egg volume (mm3) for different F2 laboratory stock populations (France 1, Bavaria 2, Sweden 3 and English 4) of Radix balthica.

2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, ,