Summary of methods used in this study.

A) Study species: the weak-biting domestic canary and the hard-biting Java sparrow. B) Schematic visualization of the biplanar high-speed X-ray videography. Marker positions were tracked in XMALab. C) Micro-CT-scans of the bird skulls and segmentation of rigid bodies, here represented on the skull of a Java sparrow. D) X-ray reconstruction of Moving Morphology (XROMM). Visualization of the anatomical coordinate system (thick axes) and the joint coordinate systems of upper and lower beaks (thin axes), here represented on the skull of a Java sparrow. For kinematic analyses, we quantified pitch angles of upper and lower beaks and yaw angles of the lower beak relative to the braincase. E) Measurements of contraction and relaxation rates of beak opener muscle (m. depressor mandibulae) and beak closer muscle (m. adductor mandibulae externus ventralis), here represented on the skull of a canary.

The tongue is critical for seed transport during positioning.

Distributions of seed locations during positioning show that the seed is moved considerably within the beak (A and B). Relationships of tongue and seed velocities (C-E) indicate that the tongue and seed move mostly together in the same direction in both species, especially along the antero-posterior axis (C). All plots show kernel density estimates with darker colors indicating more frequent occurrences. Only data of ’positioning’ phases are shown. Skull images shown in lateral and ventral views, with the lower jaw hidden in the latter. Scale bars in A and B: 5 mm. In C-E, upper right and lower left quadrants indicate movements of seed and tongue marker in the same direction along the given axis (positive correlation). Upper left and lower right quadrants indicate opposing movements along the given axis (negative correlation). Red elements in icons in C-E indicate the object (tongue or seed) and axis of the anatomical coordinate system (x: posterior-to-anterior axis, y: ventral-to-dorsal axis, z: left-to-right axis). N denotes the number of frames included for calculation of kernel density estimates. All plots show the pooled data of two individuals per species.

Birds use distinct types of seed rotation during positioning and seed orientation during biting.

A) Four dominant types of seed rotation in canaries and Java sparrows. Solid arrows in (A) indicate approximate typical trajectories of seeds during rotation, dashed arrows indicate alternative options of trajectories that the birds used only occasionally. B) Five dominant types of seed orientation during biting in canaries and Java sparrows. ’Upright’ refers to a special case of transverse orientation, with the long axis of the seed lying parallel to the sagittal plane of the beak. In A+B, ‘longitudinal’ and ‘transverse’ refer to the long axis of the seed being parallel and orthogonal to the long axis of the beak, respectively. Seed rotations and orientations are visualized on only one side for clarity. Seeds and beaks are not drawn to scale. N denotes the sample size (number of analyzed seed rotation or biting events). Wide boxes show beaks in lateral view, narrow boxes show beaks in frontal view. All plots show the pooled data of two individuals per species.

The tongue is critical for seed stabilization during biting.

Distributions of seed locations during biting indicate the consistent seed position in either of the lateral grooves of the upper palate (A and B). In both species, the tongue and seed move less compared to the positioning phase (C-E, cf. Figure 2), indicating a stabilizing function of the tongue. All plots show kernel density estimates with darker colors indicating more frequent occurrences. Only data of ’biting’ phases are shown. Skull images shown in lateral and ventral views, with the lower jaw hidden in the latter. Scale bars in A and B: 5 mm. In C-E, upper right and lower left quadrants indicate movements of seed and tongue marker in the same direction along the given axis (positive correlation). Upper left and lower right quadrants indicate opposing movements along the given axis (negative correlation). Icons in C-E indicate the object (tongue or seed) and axis of the anatomical coordinate system (x: posterior-to-anterior axis, y: ventral-to-dorsal axis, z: left-to-right axis). N denotes the number of frames included for calculation of kernel density estimates. All plots show the pooled data of two individuals per species.

Motions of both the upper and the lower beaks are involved in seed processing.

Distributions of peaks in total gape angle reveal differences between phase types and species (A). Both upper and lower beaks contribute to total gapes (B). Kernel density estimates (KDE) of yaw angles show that both species use medio-lateral rotations of the lower beak during all phases (C). Standard deviations of yaw angles of lower beak reveal differences between the species for the positioning and dehusking phase (D). Dashed lines in A and B and boxplots in D indicate the quartiles of the distributions. The whiskers in D indicate the highest and lowest data point that lies within the 1.5 inter-quartile range of the upper and lower quartile. N denotes sample sizes (A+B: number of extracted peaks in total gape angle, C: number of frames included for calculation of kernel density estimates, D: number of sequences of which yaw stdev was calculated). All plots show the pooled data of two individuals per species. See Appendix 1—table 1 and Appendix 1—table 2 for results of the statistical analyses. Source data 1. Data of maximal total gape angles (A) and associated pitch angles of upper and lower beaks (B). Source data 2. Standard deviations of lower beak yaw rotations (D).

Coordination of upper and lower beak movements during positioning, biting, and dehusking.

The narrow and diagonal density plot in the left column indicates strong negative correlation of upper and lower beaks during positioning. See main text for details. Plots show kernel density estimates of pitch angular velocities of the upper and lower beaks during positioning, biting, and dehusking. More frequent combinations of values are indicated by darker colors. N denotes the number of frames included for calculation of kernel density estimates. All plots show the pooled data of two individuals per species.

Coordination of beak movement and seed transport.

Low speeds of seed movement at low gape angles (A, see gray hatched areas) indicate that the beak needs to release its grip to allow for fast seed movements. Canaries moved the seed mostly anteriorly during beak opening and posteriorly during beak closing (B). Java sparrows showed a less specific pattern, with anterior and posterior seed transport being less strictly associated with either beak opening or closing. Plots show kernel density estimates of combinations of seed speed and gape angle (A) and seed velocity and mandible angular velocity (B). More frequent combinations of values are indicated by darker colors. N denotes the number of frames included for calculation of kernel density estimates. All plots show the pooled data of two individuals per species.

The frequency of lower beak oscillation differs between species (A) and may decrease with amplitude (B).

Plots show the lower beak pitch angle oscillation frequency during seed positioning. On average, the frequency is higher in canaries than in Java sparrows (18.6±0.4 Hz vs. 13.5±0.4 Hz, A). In Java sparrows, the frequency of oscillation decreases with increasing amplitude (B). This relationship is not significant in canaries. Dashed lines in A indicate quartiles of the distributions. Area in translucent shading in B indicates the 94% confidence interval of the regression. N denotes sample sizes, and the sample sizes per species are the same in B) as in A). Plots show the pooled data of two individuals per species. See Appendix 1—table 3A for results of the statistical analyses. Source data 1. Data of lower beak oscillation frequencies and amplitudes.

Contractile properties of adductor and depressor muscles.

A+B) Force-time curves of isometric contraction of depressor and adductor, respectively. Force data are normalized by maximal force and show the active force only (passive force has been subtracted). The gray bar indicates the duration of the tetanus stimulation (start: 0 s, end: 0.5 s). Force curves in A and B show pooled data of four individuals per species and per muscle type. C) Time to reach 50% of maximal force during isometric muscle contraction under tetanus stimulation. D) Time to reach 50% of maximal force during muscle relaxation. N denotes sample sizes (number of tetani). Violin plots show the pooled data of four individuals per species and per muscle type. Gray hatched lines indicate mean values as obtained from statistical analyses (Appendix 1—table 3B). Source data 1. Half-rise times (C) and half-relaxation times (D) of jaw muscles.

Example plots of kinematic parameters obtained from XROMM animations for a canary.

From top to bottom, the panels show the following kinematic parameters as function of time: total gape pitch angle of the beak, the pitch angle of the upper beak, the pitch and yaw angle of the lower beak, the antero-posterior translation of seed and tongue, the dorso-ventral translation of seed and tongue, the medio-lateral translation of seed and tongue, and the translation of the lower beak JCS in all three dimensions. Colored shading in the background indicates the different phase types as labeled on top of the plots. Data of parameters labeled with ’centr.’ were centered (subtraction of the mean) to allow for combined visualization of multiple parameters.

Example plots of kinematic parameters obtained from XROMM animations for a Java sparrow.

From top to bottom, the panels show the following kinematic parameters as function of time: total gape pitch angle of the beak, the pitch angle of the upper beak, the pitch and yaw angle of the lower beak, the antero-posterior translation of seed and tongue, the dorso-ventral translation of seed and tongue, the medio-lateral translation of seed and tongue, and the translation of the lower beak JCS in all three dimensions. Colored shading in the background indicates the different phase types as labeled on top of the plots. Data of parameters labeled with ’centr.’ were centered (subtraction of the mean) to allow for combined visualization of multiple parameters.

Tongue velocity during seed positioning in canaries (A) and Java sparrows (B).

During seed rotation, the tongue moves faster in canaries (see black arrows in A), but much less so in Java sparrows (B). Plots show kernel density estimates of antero-posterior velocity of the tongue as function of total gape angle. More frequent combinations are indicated by darker colors. Solid lines indicate linear regressions of the data. Plots show the pooled data of two individuals per species.

Results of the statistical analysis for maxima in total pitch angle (cf. Figure 5A) and the corresponding pitch angles of upper and lower beak (cf. Figure 5B) for canaries (’Can’) and Java sparrows (’Jav’).

Data are shown for positioning (’Pos.’), biting (’Bit.’), and dehusking (’Deh.’). Values for mean, standard deviation (stdev), and the lower and upper bounds of the highest density interval (hdi3% and hdi97%) are derived from the posterior distributions of the Bayesian statistical model (see Methods for details). The rows with the label ’Can-Jav’ indicate the differences between the tested species.

Results of the statistical analysis of medio-lateral movements (yaw angles) of the lower beak (cf. Figure 5D) during positioning (’Pos.’), biting (’Bit.’), and dehusking (’Deh.’).

Labels ’Can.’ and ’Jav.’ refer to species names ’canary’ and ’Java sparrow’, respectively. Labels ’Can.-Jav.’ refer to differences between species. The standard deviation (’stdev’) of the distribution of used yaw angles was used to quantify the width of range of lower beak yaw rotations. Values for mean, stdev, and the lower and upper bounds of the highest density interval (hdi3% and hdi97%) are derived from the posterior distributions of the Bayesian statistical model (see Methods for details).

Results of the statistical analysis of A) the oscillation frequency of the lower beak (cf. Figure 8) and B) the rise and relaxation times of adductor and depressor muscles (cf. Figure 9).

Labels ’Can’ and ’Jav’ refer to species names ’canary’ and ’Java sparrow’, respectively. Labels ’depr’ and ’add’ refer to the muscle types ’depressor’ and ’adductor’, respectively. Values for mean, standard deviation (stdev) and the lower and upper bounds of the highest density interval (hdi3% and hdi97%) are derived from the posterior distributions of the Bayesian statistical model (see Methods for details). Labels ’Can.-Jav.’ and ’depr.-add.’ refer to differences between species or muscle types, respectively.