Research ArticleEVOLUTIONARY BIOLOGY

Structure-function covariation with nonfeeding ecological variables influences evolution of feeding specialization in Carnivora

See allHide authors and affiliations

Science Advances  07 Feb 2018:
Vol. 4, no. 2, eaao5441
DOI: 10.1126/sciadv.aao5441
  • Fig. 1 Cranial shape deviations associated with ecological traits.

    Background arrows indicate direction of change depicted. Heat maps show range of negative deviations (cool colors) to positive deviations (hot colors). Deviation heat maps and shape deformations were performed on the consensus cranial shape.

  • Fig. 2 Engineering-based performance measures analyzed in cranial models.

    (A) Stiffness (strain energy density) in bending tests. (B) Safety factor [von Mises (VM) stress] in bending tests. (C) Stiffness in temporalis-driven tests. (D) VM stress in temporalis-driven tests. (E) Mechanical efficiency in temporalis-driven tests. (F) Stiffness in masseter-driven tests. (G) VM stress in masseter-driven tests. (H) Mechanical efficiency in masseter-driven tests. Error bars indicate 95% CI; if no error bars are visible, error is encompassed within the size of the mean value dots. Mpa, megapascals.

  • Fig. 3 Summary figure with directions for future research.

    A flowchart represents the relationship between the environmental, life history, ecological, shape, and feeding performance variables analyzed in this study. Asterisks denote areas that might especially benefit from additional research to understand how life history and environmental variation directly influence cranial shape variation. Shaded arrows indicate likely feedback pathways that were not specifically analyzed in this study but, nevertheless, are important for future research into the relationships among feeding ecology, biomechanical performance, and cranial shape.

  • Table 1 Statistical significance (P values) of PGLS regressions based on resampling permutations (n = 1000).

    Statistically significant (at the P = 0.05 level) regressions shown in boldfaced font. For combined regression analyses of multiple variables, none are statistically significant; P values of the individual factors are shown in those cases. For sensitivity analyses, see table S4.

    CarnivoraCaniformiaFeliformia
    Number of species533122
    GMM data (G)
      G1. CS0.030.170.04
    Locomotion and activity (LA)
      LA1. Activity cycle0.360.600.67
      LA2. Terrestriality/arboreality0.620.470.61
      LA3. Habitat breadth0.640.660.61
      LA1 + 2 + 30.44/0.76/0.760.65/0.79/0.420.68/0.93/0.86
    Feeding ecology (FE)
      FE1. Dietary breadth0.010.310.003
      FE2. Trophic level0.030.080.03
      FE1 + 20.03/0.030.40/0.130.01/0.04
    Life history (LH)
      LH1. Longevity0.180.240.36
      LH2. Age at sexual maturity0.020.030.65
      LH1 + LH20.40/0.060.29/0.130.86/0.77
    Environmental variables (EV)
      EV1. Mean precipitation0.060.390.08
      EV2. Mean temperature0.450.440.38
      EV1 + 20.07/0.350.43/0.410.14/0.84
  • Table 2 Shape changes by cranial region in each of the ecological gradients significantly associated with cranial shape covariation.

    Category abbreviations as in Table 1.

    RostrumPostorbital constrictionCranial vaultOccipital regionZygomatic arches
    G1. Small to largeDeeperExpandedNarrowerExpandedNo change
    FE1. Broad to narrowShorterExpandedWiderNo changeWider
    LH2. Early to lateDeeperNarrowerWiderExpandedNo change
    EV1. Low to highLongerNarrowerNarrowerExpandedNarrower
    FE2. Consensus (herbivory)ShorterNarrowerTallerRetractedWider
    FE2. Consensus (omnivory)Longer + deeperNarrowerWiderExpandedNarrower
    FE2. Consensus (carnivory)ShorterExpandedNarrowerNo ChangeWider
  • Table 3 Results of linear regression model fits of ecological variables that covary significantly with cranial shape, against other ecological variables.

    df, degrees of freedom; Adj. SS, adjusted sum of squares; Adj. MS, adjusted mean square; F, F statistic; P, significance level.

    dfAdj. SSAdj. MSFPDirection of
    change
    CS
      Regression132,484,639191,1263.620.004*
      Mean monthly
    temperature
    1127,563127,5632.420.135
      Sexual maturity
    age
    1709,585709,58513.450.001*Later
      Maximum
    longevity
    146,66446,6640.880.358
      Mean monthly
    precipitation
    134,24434,2440.650.429
      Activity cycle262,30831,1540.590.563
      Terrestriality/
    arboreality
    130,67230,6720.580.454
      Habitat breadth241,33020,6650.390.681
      Diet breadth4196,92649,2310.930.464
      Error211,107,57852,742
      Total343,592,217
    Sexual maturity age
      Regression124,301,175358,4314.140.002*
      Mean monthly
    temperature
    116,199161,990.190.669
      Maximum
    longevity
    11,488,5251,488,52517.21<0.001*Longer
      Mean monthly
    precipitation
    111,43411,4340.130.72
      Activity cycle2262,025131,0121.510.242
      Terrestriality/
    arboreality
    126,78526,7850.310.584
      Habitat breadth2268,973134,4871.550.234
      Diet breadth41,518,075379,5194.390.009*Pandas
      Error221,902,97286,499
      Total346,204,148
    Precipitation
      Regression1242,736.33,561.363.740.004*
      Mean monthly
    temperature
    18,3418,3418.770.007*Higher
      Sexual maturity
    age
    1125.8125.770.130.72
      Maximum
    longevity
    147.647.590.050.825
      Activity cycle2520.3260.130.270.763
      Terrestriality/
    arboreality
    15,226.95,226.945.490.029*Aboveground
      Habitat breadth2157.178.540.080.921
      Diet breadth46,525.31,631.331.710.183
      Error2220,933.4951.52
      Total3463,669.7

    *Statistical significance at the P = 0.05 level.

    • Table 4 Engineering-based performance measure shifts associated with ecological traits.

      Category abbreviations as in Table 1.

      Temporalis-drivenMasseter-driven
      StiffnessSafety factorEfficiencyStiffnessSafety factorEfficiency
      G1. Small to largeNo changeHigherLowerNo changeHigherHigher
      FE1. Broad to narrowHigherHigherHigherLowerNo changeNo change
      LH2. Late to earlyHigherHigherLowerLowerNo changeNo change
      EV1. High to lowHigherHigherHigherHigherHigherHigher
      FE2. Omnivory-herbivoryHigherHigherLowerHigherHigherHigher
      FE2. Carnivory-herbivoryHigherHigherLowerHigherHigherLower
    • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/2/eaao5441/DC1

      Additional Methodological Details and Results

      table S1. List of specimens used in the study.

      table S2. List of anatomical landmarks and semilandmarks used in the GMM analyses.

      table S3. Ecological attributes of species analyzed.

      table S4. Summary of sensitivity analyses of PGLS regression using different phylogenetic tree topology and branch length configurations.

      table S5. Area and volume attributes of the theoretical morphed models analyzed.

      table S6. Output values of FE simulations in the “bending test” analysis.

      table S7. Output values of FE simulations in the “temporalis-driven” analysis.

      table S8. Output values of FE simulations in the “masseter-driven” analysis.

      table S9. List of the highest mechanical demands from food items experienced by each carnivoran species, ranked on the basis of stiffness and density.

      fig. S1. Anatomical landmarks used in the GMM analyses.

      fig. S2. Main analysis.

      fig. S3. Sensitivity analysis 1.

      fig. S4. Sensitivity analysis 2.

      fig. S5. Sensitivity analysis 3.

      fig. S6. Sensitivity analysis 4.

      fig. S7. Sensitivity analysis 5.

      fig. S8. Sensitivity analysis 6.

      fig. S9. Sensitivity analysis 7.

      fig. S10. Sensitivity analysis 8.

      fig. S11. Sensitivity analysis 9.

      fig. S12. Contour plot of the first two PC axes against life history and environmental variables.

      fig. S13. 3D plots of PC scores of cranial shape data.

      fig. S14. Plot of mean annual precipitation versus mean annual temperature of species studied.

      References (6972)

    • Supplementary Materials

      This PDF file includes:

      • Additional Methodological Details and Results
      • table S1. List of specimens used in the study.
      • table S2. List of anatomical landmarks and semilandmarks used in the GMM analyses.
      • table S3. Ecological attributes of species analyzed.
      • table S4. Summary of sensitivity analyses of PGLS regression using different phylogenetic tree topology and branch length configurations.
      • table S5. Area and volume attributes of the theoretical morphed models analyzed.
      • table S6. Output values of FE simulations in the “bending test” analysis.
      • table S7. Output values of FE simulations in the “temporalis-driven” analysis.
      • table S8. Output values of FE simulations in the “masseter-driven” analysis.
      • table S9. List of the highest mechanical demands from food items experienced by each carnivoran species, ranked on the basis of stiffness and density.
      • fig. S1. Anatomical landmarks used in the GMM analyses.
      • fig. S2. Main analysis.
      • fig. S3. Sensitivity analysis 1.
      • fig. S4. Sensitivity analysis 2.
      • fig. S5. Sensitivity analysis 3.
      • fig. S6. Sensitivity analysis 4.
      • fig. S7. Sensitivity analysis 5.
      • fig. S8. Sensitivity analysis 6.
      • fig. S9. Sensitivity analysis 7.
      • fig. S10. Sensitivity analysis 8.
      • fig. S11. Sensitivity analysis 9.
      • fig. S12. Contour plot of the first two PC axes against life history and environmental variables.
      • fig. S13. 3D plots of PC scores of cranial shape data.
      • fig. S14. Plot of mean annual precipitation versus mean annual temperature of species studied.
      • References (69–72)

      Download PDF

      Files in this Data Supplement:

    Navigate This Article