Growth and differentiation during delayed metamorphosis of feeding gastropod larvae: signatures of ancestry and innovation

Extent of larval growth among marine invertebrates has potentially profound implications for performance by benthic recruits because body size influences many biological processes. Among gastropods, feeding larvae often attain larger size at metamorphic competence than non-feeding larvae of basal gastropod clades. Delay of metamorphosis can further influence size at recruitment if larvae continue to grow during the delay. Some caenogastopod larvae grow during delayed metamorphosis, but opisthobranch larvae do not. Data on larval growth of neritimorph gastropods are needed to help determine which of these growth patterns for planktotrophic gastropod larvae is more derived. We cultured planktotrophic larvae from all three major gastropod clades with feeding larvae through delays of metamorphosis of 3–10 weeks. Larvae of the caenogastropod Euspira lewisii and the euthyneurans Haminoea vesicula (Opisthobranchia) and Siphonaria denticulata (Pulmonata) conformed to previously described growth patterns for their respective major clades. Furthermore, the caenogastropod continued to lengthen the prototroch (ciliary band for swimming and feeding) and to differentiate prospective post-metamorphic structures (gill filaments and radular teeth) during delayed metamorphosis. Larvae of the neritimorph Nerita atramentosa arrested shell growth during delayed metamorphosis but the radula continued to elongate, a pattern most similar to that of non-feeding larvae of Haliotis, a vetigastropod genus. Character mapping on a phylogenetic hypothesis suggests that large larval size and capacity for continued growth during delayed metamorphosis, as exhibited by some caenogastropods, is a derived innovation among feeding gastropod larvae. This novelty may have facilitated post-metamorphic evolution of predatory feeding using a long proboscis.     

Rearing Pandalus borealis larvae in the laboratory

Larvae of the northern shrimp Pandalus borealis (Krøyer) are pelagic. In the Estuary and Gulf of St. Lawrence, Canada, the early stages are found in the upper 25-m of the water column in spring and early summer and are expected to experience a range of water temperatures from as low as 0°C to as high at 6°C. Little is known of the impact of water temperature on metabolic requirements of northern shrimp larvae. In this study, routine respiration (VO₂), maximum respiration (electron transport system activity, ETSA) and metabolic scope for growth (MS, ETSA–VO₂) of northern shrimp larvae were measured as a function of temperature (3, 5 and 8°C), developmental stage (I–V at 3°C, I–VII at 5°C and 8°C) and growth rate in dry mass. After logarithmic transformation, all three metabolic variables were linearly related to dry mass. The increase in VO₂ with body mass was faster at 5°C than at 3 or 8°C, whereas with ETSA this increase was slower. As a result, MS increased more slowly with dry mass at 5°C than at 3 and 8°C. However, MS did not limit growth in this study, since it explained only 39% of the variability in growth. All three metabolic variables as well as growth varied together as a function of temperature and ontogeny. Q₁₀ of all three metabolic variables ranged from 1.6 and 2.2 for stages I–V larvae, except for VO₂ at stage I (3.9) and stage III (2.9).     

Induction of metamorphosis in larvae of the oyster Crassostrea gigas using neuroactive compounds

Chemical (neuroactive compounds at different concentrations and exposure times) and physical (water agitation, light) factors with potential effects on the metamorphosis of larvae of the oyster Crassostrea gigas (Thunberg) larvae have been studied. The neurotransmitters l-dihydroxyphenylalanine (DO), epinephrine (EP), norepinephrine (NE), and acetylcholine (AC) have been identified as very active inducers of metamorphosis, whilst serotonin (SE), dopamine (DA) and potassium (K) were less effective inducers. The -aminobutyric acid (GA) and ammonium (AM) were found ineffective at the concentrations tested. Exposure to 10^-4 M EP for 15 min was sufficient to promote >80% metamorphosis within 48 h, whereas NE required 2 h to exert comparable induction. Maximum induction by DO (>50%) was achieved after 2 h exposure to 10^-4 M. However, unlike EP and NE, DO was lethal at that concentration in the long term. Maximum induction by AC (30% metamorphosis) was achieved at a concentration of 10^-4 M. In contrast to other neurotransmitters, AC induced settlement behaviour, cementation and eventual metamorphosis, yielding postlarvae which were all attached to the substratum. EP and NE triggered the morphogenetic changes, by-passing settlement and leading to postlarvae not cemented to the substratum. DO induced mostly attached spat at low concentrations (10^-5 M) and unattached spat at high concentrations (10^-4 M), and a similar pattern was apparent for the weaker inducers SE and DA. Regarding physical factors, a highly reflectant surface significantly increased the percentage of attached spat obtained, compared to a dark bottom. No consistent effect of water current or light was detected on the production of unattached spat. The three different forms of induction are discussed in relation to different regulatory pathways of settlement and metamorphosis.     

Dissolved histamine: a potential habitat marker promoting settlement and metamorphosis in sea urchin larvae

Many species of marine invertebrate larvae settle and metamorphose in response to chemicals produced by organisms associated with the adult habitat, and histamine is a cue for larvae of the sea urchin Holopneustes purpurascens. This study investigated the effect of histamine on larval metamorphosis of six sea urchin species. Histamine induced metamorphosis in larvae of three lecithotrophic species (H. purpurascens, Holopneustes inflatus and Heliocidaris erythrogramma) and in one planktotrophic species (Centrostephanus rodgersii). Direct comparisons of metamorphic rates of lecithotrophic and planktotrophic larvae in assays cannot be made due to different proportions of larvae being competent. Histamine (10 μM) induced metamorphosis in 95% of larvae of H. purpurascens and H. inflatus after 1 h, while the coralline alga Amphiroa anceps induced metamorphosis in 40–50% of these larvae. Histamine (10 μM) and A. anceps induced 40 and 80% metamorphosis, respectively, in the larvae of H. erythrogramma after 24 h. Histamine (10 μM) and the coralline alga Corallina sp. induced 30 and 70% metamorphosis, respectively, in the larvae of C. rodgersii after 24 h. No metamorphosis of any larval species occurred in seawater controls. Larvae of two planktotrophic species (Tripneustes gratilla and Heliocidaris tuberculata) did not metamorphose in response to histamine. Seagrasses, the host plants of H. inflatus, induced rapid metamorphosis in larvae of the two Holopneustes species, and several algae induced metamorphosis in C. rodgersii larvae. Histamine leaching from algae and seagrasses may act as a habitat marker and metamorphic cue for larvae of several ecologically important sea urchin species.     

Settlement behavior and metamorphosis of oyster larvae (Crassostrea gigas) in response to bacterial supernatants

Veliger larvae of the oysterCrassostrea gigas (Thunberg) responded to unknown dissolved chemical inducers found in supernatants of cultures of the bacteriaAlteromonas colwelliana andVibrio cholerae. The response, which was similar to that seen when larvae were exposed to the neurotransmitter precursor L-3,4-dihydroxyphenylalanine (L-DOPA), consisted of an initial settlement phase of swimming with the foot extended and crawling on the substrate. Subsequently larvae attached to the substrate and metamorphosed. The percentage of veligers metamorphosing following inducation of settlement behavior was higher in a group of older larvae, a response similar to that seen with L-DOPA, suggesting that competence to respond to bacterial supernatants is divided into two phases: behavioral competence followed by morphogenetic competence. Following size exclusion chromatography, the molecular weight of the peak containing the activity which induced settlement behavior was determined to be 300 daltons. Autoclaved Marine Broth, which induced low levels of settlement behavior also contained this low molecular weight active peak, suggesting that an oyster settlement inducer is also present in this medium.Contribution # 137 from the Center of Biotechnology, Marine Biotechnology Institute, University of Maryland, USA     

Diel vertical migration and timing of metamorphosis of larvae of the Dungeness crab Cancer magister

Joint USA/USSR ichthyoplankton surveys off the coasts of Washington, Oregon and northern California during the years 1981 to 1985 sampled more than 120 stations each year, from 5 to 360 km offshore and between Latitude 40° and 48° N, providing information on ontogeny and diel migration of larvae of the Dungeness crab Cancer magister on a scale not studied previously. We developed a maximum likelihood method for estimating abundance and fraction in the neuston at each station from a neuston tow and an oblique bongo tow. Latestage megalopae migrate vertically on a diel basis, with the fraction in the neuston being (on average) 62% at night (19.00 to 08.00 hrs Pacific Standard Time, PST) and 8% during the day (08.00 to 19.00 hrs PST). The hourly pattern of this migration includes a peak in the early evening, possibly another in the early morning, and an intermediate level in the late afternoon. We detected no dependence of vertical migration on cloud cover or sea state. Early-stage megalopae were present in much lower fractions in the neuston, but weakly displayed the same diel pattern of migration. Zoeae appeared to be below the neuston at all times, except for 2 or 3 h in the evening. From an abrupt change in larval stage in samples from a north-south cruise, we concluded that the majority of the larvae metamorphose from zoeae to megalopae over a fairly short time span (2 to 4 wk) at a given latitude. In later cruises, 95% of the larvae were megalopae, indicating that metamorphosis over the study area either occurs at the same time or proceeds from south to north over a time span of less than a month in early spring.     

Development and metamorphosis of the planktotrophic larvae of Rostanga pulchra (Mollusca: Nudibranchia)

Rostanga pulchra MacFarland, a small (1 to 2 cm) dorid nudibranch, lays an average of 7000 eggs in the laboratory during a period of 30 days in the summer. The veligers hatch 15 to 16 days after oviposition and it takes another 35 to 40 days to become competent for metamorphosis at a temperature of 10° to 15°C. Larval cultures were maintained initially at a concentration of 500 veligers per 100 ml of filtered sea water (antibiotics added). During the planktotrophic phase of development, the veliger grows from 150 to 300 m in shell length. Although the veligers are generalists in their food preference, the best result (faster growth) was achieved by feeding them with a combination of Monochrysis lutheri and Isochrysis galbana. The concentration of food cells was kept at 10⁴ cells per ml of culture media and was supplied every 2 to 3 days. A veliger which is competent to metamorphose is identifiable morphologically by its propodium, eyespots, rhinophores, and spiculated dorsal papillae. The entire metamorphic process lasts 24 h when a suitable substrate such as the food sponge Ophlitaspongia pennata is provided. The competent veliger is able to delay metamorphosis for at least 3 weeks. Juveniles were kept in the laboratory for 70 days and, during this period, grew to a length of 4.5 mm.     

A method for rearing Pecten maximus larvae in the laboratory

Development of the scallop Pecten maximus (L.) from egg to metamorphosis takes 33 to 38 days at 16°C. The shelled veliger first appears 2 days after fertilisation, and crawling pediveligers at around 2 days before metamorphosis. The pediveliger can attach temporarily to filamentous objects by means of a byssus thread and, even after metamorphosis, periods of attachment alternate with periods of crawling. The larvae are reared in polythene bins containing 30 l of full salinity sea water (pre-filtered through 0.2 filters), which is changed every 2 days. Mixtures of algal foods (Isochrysis galbana, Chaetoceros calcitrans, Pyramimonas obovata and Tetraselmis suecica), concentrated by centrifugation, are added at each change from the third day onwards to give a concentration of 50 cells/l. Larval density is maintained below 10/ml. Antibiotics are added at each change, but very careful sterilisation of all equipment is necessary to guard against the introduction of diseases.     

Food limitation stimulates metamorphosis of competent larvae and alters postmetamorphic growth rate in the marine prosobranch gastropodCrepidula fornicata

The effects of food limitation on growth rates and survival of marine invertebrate larvae have been studied for many years. Far less is known about how food limitation during the larval stage influences length of larval life or postmetamorphic performance. This paper documents the effects of food limitation during larval development (1) on how long the larvae ofCrepidula fornicata (L.) can delay metamorphosis in the laboratory after they have become competent to metamorphose and (2) on postmetamorphic growth rate. To assess the magnitude of nutritional stress imposed by different food concentrations, we measured growth rates (as changes in shell length and ash-free dry weight) for larvae reared in either 0.45-m filtered seawater or at phytoplankton concentrations (Isoehrysis galbana, clone T-ISO) of 1 × l0³, 1 × 10⁴, or 1.8 × 10⁵ cells ml^–1. Larvae increased both shell length and biomass at 1 × 10⁴ cells ml^–1, although significantly more slowly than at the highest food concentration. Larvae did not significantly increase (p > 0.10) mean shell length in filtered seawater or at a phytoplankton concentration of only 1 × 10³ cells ml^–1, and in fact lost weight under these conditions. To assess the influence of food limitation on the ability of competent individuals to postpone metamorphosis, larvae were first reared to metamorphic competence on a high food concentration ofI. galbana (1.8 × 10⁵ cells ml^–1). When at least 80% of subsampled larvae were competent to metamorphose, as assessed by the numbers of indlviduals metamorphosing in response to elevated K⁺ concentration in seawater, remaining larvae were transferred either to 0.45-m filtered seawater or to suspensions of reduced phytoplankton concentration (1 × 10³, 1 × 10⁴, or 5 × 10⁴ cells ml^–1), or were maintained at 1.8 × 10⁵ cells ml^–1. All larvae were monitored daily for metamorphosis. Individuals that metamorphosed in each food treatment were transferred to high ration conditions (1.8 × 10⁵ tells ml^–1) for four additional days to monitor postmetamorphic growth. Competent larvae responded to all food-limiting conditions by metamorphosing precociously, typically 1 wk or more before larvae metamorphosed when maintained at the highest food ration. Surprisingly, juveniles reared at full ration grew more slowly if they had spent 2 or 3 d under food-limiting conditions as competent larvae. The data show that a rapid decline in phytoplankton concentration during the larval development ofC. fornicata stimulates metamorphosis, foreshortening the larval dispersal period, and may also reduce the ability of postmetamorphic individuals to grow rapidly even when food concentrations increase.     

Growth and elemental composition (C,N, H) in Inachus dorsettensis (Decapoda: Majidae) larvae reared in the laboratory

Larvae of the spider crab Inachus dorsettensis were reared in the laboratory at constant 12 °C. Development lasted 8 to 10 d in the Zoea I, 10 to 12 d in the Zoea II and 14 to 20 d in the megalopa stage. During this time, larval growth was measured in samples taken every 2 to 4 d as dry wt (W), carbon (C), nitrogen (N), and hydrogen (H); energy content (E) was calculated from C. Biomass and energy (per individual) increased in each larval stage as a parabola-shaped function of age, which could be fitted by a power equation. C, H, and E show a higher percentage gain (relative to the initial values at hatching) than W or N, suggesting that proportionally more lipid than protein is accumulated during larval development. There are cyclical changes in the relative (per unit of W) biomass and energy figures, corresponding to the larval moult cycles: immediately after each ecdysis all these values decrease, presumably due to rapid uptake of water and minerals, then they increase again due to tissue growth and remain high until the next moulting occurs. Cyclical changes in the C/N ratio suggest that proportionally more lipid than protein is accumulated during the initial (postmoult) phase of the moult cycle, followed by a period of balanced or protein-dominated gain during the intermoult and premoult phases. These patterns of growth and elemental composition observed during the complete larval development and in single moult cycles of I. dorsettensis are compared with those described in the literature for other decapod species. This comparison suggests a high degree of similarity in biochemical composition and growth characteristics of larval decapod crustaceans.     

The life history of Rhodophysema georgii in laboratory culture

Rhodophysema georgii Batt., a rare epiphyte on Zostera spp., was collected from lowtide-level lagoons at Bembridge (Isle of Wight), and its life history examined in laboratory culture. The tetraspores, the only method of reproduction known, gave rise to small crusts which again developed tetrasporangial sori on the surface, similar to those of the parental material. It is suggested that this direct type of life history is probably due to the repression of meiosis in the tetrasporangial mother cell.     

Feeding,growth, and survival of Engraulis mordax larvae reared in the laboratory

Methods are described for the successful rearing of northern anchovy larvae (Engraulis mordax Girard) on cultured foods. Larvae were fed successively on the unarmored dinoflagellate Gymnodinium splendens, the veliger of the gastropod Bulla gouldiana, and nauplii of the brine shrimp Artemia salina. Rearing containers ranging in capacity from 4.5 to 510 l were tested; the smaller ones were found to be most useful for laboratory experimentation. Irreversible starvation occurred when E. mordax were denied food for more than 1.5 days after yolk absorption. Growth rates of larval anchovies fed different diets were compared. Larvae fed G. splendens grew for 1 week at the same rate as animals fed wild plankton, but did not maintain this rate. Laboratory survival of E. mordax larvae on a diet of G. splendens alone, did not differ significantly when veligers supplemented the diet. However, when G. splendens and veligers were fed simultaneously to E. mordax larvae, growth rate was greatly improved, although still not matching the growth attained on a diet of wild plankton. Length (L) versus weight (W) analyses were made for all larvae at all diets. The results showed that weight could be calculated most accurately from length by the relationship log W=3.3237 log L-3.8205, regardless of diet.     

Nutrient uptake kinetics and growth of zooxanthellae maintained in laboratory culture

Growth characteristics and nutrient uptake kinetics were determined for zooxanthellae (Gymnodinium microadriaticum) in laboratory culture. The maximum specific growth rate (_max) was 0.35 d^-1 at 27 °C, 12 hL:12 hD cycle, 45 E m^-2 s^-1. Anmmonium and nitrate uptake by G. microadriaticum in distinct growth phases exhibited Michaelis-Menten kinetics. Ammonium half-saturation constants (K_s) ranged from 0.4 to 2.0 M; those for nitrate ranged from 0.5 to 0.8 M. Ammonium maximum specific uptake rates (V_max) (0.75 to 1.74 d^-1) exceeded those for nitrate (0.14 to 0.39 d^-1) and were much greater than the maximum specific growth rate (0.35 d^-1), suggesting that ammonium is the more significant N source for cultured zooxanthellae. Ammonium and nitrate V_max values compare with those reported from freshly isolated zooxanthellae. Light enhanced ammonium and nitrate uptake; ammonium inhibited nitrate uptake which was not reported for freshly isolated zooxanthellae, suggesting that physiological differences exist between the two. Knowledge of growth and nutrient uptake kinetics for cultured zooxanthellae can provide insight into the mechanisms whereby nutrients are taken up in coral-zooxanthelae symbioses.Contribution No. 1515 from the University of Maryland Center for Environmental and Estuarine Studies, Chesapeake Biological Laboratory, Solomons, Maryland 20688-0038, USA     

Requirement of exogenous inducers for metamorphosis of axenic larvae and buds of Cassiopea andromeda (Cnidaria: Scyphozoa)

Planula larvae and asexually-produced buds of the rhizostome scyphozoan Cassiopea andromeda (collected throughout the year in Eilat, Israel) have the ability, under axenic conditions, to attach to a substrate and undergo morphogenetic development to form a polyp (=scyphistoma) in: (1) the presence of unidentified inducers found in the adult habitat and (2) the presence of cefined organic compounds. Axenic planulae and buds were unable to settle and complete metamorphosis in autoclaved artificial or natural seawater from the North Sea when maintained without food, but continued swimming while decreasing in size and protein content, eventually dying within three months. When maintained in autoclaved seawater from the Red Sea, between 25 and 46% of the planulae and 4 and 11% of the buds metamorphosed within 30 d. Axenic solutions of cholera toxin, thyroid stimulating hormone, and pancreatic casein hydrolysate peptides in artificial seawater induced morphogenic development of 20 to 100% of planulae and buds within 2 to 18 d. The natural inducer(s) in Red Sea seawater, though unidentified, may have characteristics similar to the large proteins and small peptide inducers used in this study. Planulae and buds older than 20 d metamorphosed sooner and responded to lower concentrations of pancreatic casein hydrolysate peptides than younger individuals. This may be a physiological mechanism for enhancing metamorphosis and survival in nature. The data show that settlement and metamorphosis can be induced by solutions of cholera toxin and thyroid stimulating hormone, suggesting that, as in mammalian systems, the mechanism of action of these chemicals may involve cyclic adenosine monophosphate (cAMP) as an intermediate messenger. However, dibutyric cAMP, which is capable of passing through membranes and functioning normally inside the cell, did not induce metamorphosis of buds, and the levels of intracellular cAMP in buds and larvae typically increased slowly during induction of metamorphosis, unlike the high and rapid increases associated with cAMP-mediated biochemical events in mammalian cells. These results suggest that the observed cAMP changes seen were associated with metamorphic development, but not with the triggering mechanism.     

Growth maximization in early sardine larvae: a metabolic approach

Sardine larvae are forced to grow as fast as possible to reduce larval mortality. Thus, changes in biochemical composition during the first steps of sardine larval growth are intended to maximize larval growth rate efficiency and survival. Protein and RNA weight-specific growth rates were the highest and their corresponding doubling times the shortest among all the biomolecules, reflecting the importance of fast growth during early stages of larval development. The protein percentage increased and the carbohydrate and lipid percentages decreased during early growth until they reached, respectively, a percentage of 73.7, 3.1 and 18.0%. These percentages would represent the optimal proportion of biochemical components in sardine early larvae and they are the result of the trade-off between, in the short term, the protein proportion necessary to optimize larval movement and growth and, in the long term, the minimum lipid percentage necessary to guarantee energy reserves to fuel metamorphosis. RNA/DNA ratio increases during larval growth up to an asymptotic optimal value of ≈3.5 in postflexion larvae. Nutritional condition of sardine larvae was good and was influenced by the parental effect through the egg biochemical composition and by the growth trajectory determined by the actual environmental conditions.     

Studies on the life history of Rhodophysema elegans in laboratory culture

The life history of the red alga Rhodophysema elegans (J. Ag.) Dixon (Peyssoneliaceae: Cryptonemiales), which was commonly found at 4 localities on the south coast of England, has been investigated in laboratory culture. Tetraspores (the only reproductive organs observed) which were derived from field material germinated to repeat the parental tetrasporangiate phase. Sexual reproductive organs were not observed. It is suggested that the probable cause of this direct type of life history is the occurrence of apomeiosis in the sporangial primordium.     

A laboratory energy balance for the larvae and juveniles of the American lobster Homarus americanus

A laboratory energy budget was constructed for the larvae and juveniles of the American lobster Homarus americanus Milne-Edwards fed brine shrimp, Artemia saline L. Measured energy flows included ingestion, egestion, excretion of ammonia, routine and fed metabolism, growth, and production of exuvia. Digestion and assimilation were calculated and minimum ration of protein necessary to sustain larval lobsters was estimated. No change associated with metamorphosis was observed in rates of excretion, fed metabolism, and production of exuvia. Routine metabolism is not significantly higher for larvae than for juveniles. Growth changes from exponential in larvae to a slower increase in post-larvae. Consumption reflects changes in other variables. Changes in energy partitioning and energetic efficiencies associated with metamorphosis are largely due to change in rate of growth.     

Growth of zooxanthellae in culture with two nitrogen sources

Physiological characteristics of zooxanthellae were examined under nutrient-saturated conditions created by mixing ammonium (¹⁵NH₄) with nitrate (¹⁵NO₃) to give 0.88 mM total nitrogen. Growth rate varied with the form of nitrogen provided. Ammonium alone resulted in the lowest C:N and C:chl-a ratios. Although zooxanthellae took up nitrate in the absence of ammonium, ammonium assimilation was 1.3 times higher than nitrate assimilation. Ammonium strongly inhibited nitrate assimilation. While high-ammonium treatments resulted in the highest ¹⁴C incorporation into intermediate compounds, high nitrate levels resulted in the highest ¹⁴C incorporation into protein, suggesting that the intermediate compounds are produced prior to the subsequent production of protein when ammonium is the dominant N source. The enhanced production of intermediate compounds at the expense of carbon directed to protein synthesis in the presence of ammonium might be analogous to the “host factor” observed in zooxanthellae–host symbioses, since growth rate is depressed due to low production of protein. Received: 16 March 2000 / Accepted: 26 August 2000     

A copula-based inference to piecewise exponential models under dependent censoring,with application to time to metamorphosis of salamander larvae

In ecology and evolutionary biology, controlled animal experiments are often conducted to measure time to metamorphosis which is possibly censored by the competing risk of death and the follow-up end. This paper considers the problem of estimating the survival function of time-to-event when it is subject to dependent censoring. When the censorship is due to competing risks, the traditional assumption of independent censorship may not be satisfied, and hence, the usual application of the Kaplan–Meier estimator yields a biased estimation for the survival function of the event time. This paper follows an assumed copula approach (Zheng and Klein in Biometrika 82(1):127–138, 1995) to adjust for dependence between the event time of interest and the competing event time. While the literature on an assumed copula approach has mostly focused on semiparametric settings, we alternatively consider a parametric approach with piecewise exponential models for fitting the survival function. We develop maximum likelihood estimation under the piecewise exponential models with an assumed copula. A goodness-of-fit procedure is also developed, which touches upon the identifiability issue of the copula. We conduct simulations to examine the performance of the proposed method and compare it with an existing semiparametric method. The method is applied to real data analysis on time to metamorphosis for salamander larvae living in Hokkaido, Japan (Michimae et al. in Evol Ecol Res 16:617–629, 2014).