Cranial kinematics and prey-type effects in Amia ocellicauda feeding strikes

Whitlow, Katrina 1 ; Ross, Callum 2 ; Westneat, Mark 2

Published Feb 12, 2025; Updated Mar 24, 2025 on Dryad. https://doi.org/10.5061/dryad.c866t1ghn

Data files

Feb 12, 2025 version files 11.87 MB

Mar 17, 2025 version files 11.87 MB

Mar 24, 2025 version files 11.87 MB

Abstract

Variability in the biomechanics and kinematics of prey capture in vertebrates has been studied extensively, with evidence of multiple strategies for successful feeding in many taxa. Early research into suction feeding strikes in fishes hypothesized that fish utilize a set of pre-programmed strike kinematics that cannot be altered once initiated. However, more recent evidence has demonstrated that teleost fishes not only deploy unique strike kinematics for different prey types, but that they also alter their kinematics in response to a prey item attempting to escape. It has not yet been explicitly investigated whether non-teleost actinopterygians can also modulate the strike in response to different prey types. Here we examined the kinematics of suction strikes in bowfin, Amia ocellicauda, a holostean fish most closely related to gars. We recorded Amia feeding on both feeder fish and worms, two types of live prey differing in evasiveness, using X-Ray Reconstruction of Moving Morphology (XROMM). We found significant prey type effects on the magnitude, timing, and velocity of jaw opening, hyoid arch depression, and pectoral girdle motions. These prey type effects demonstrate that the ability to modulate feeding strikes evolved early in actinopterygian fishes, and is possibly the ancestral state for jawed vertebrates.

https://doi.org/10.5061/dryad.c866t1ghn

Description of the data and file structure

Files and variables

File: VolDataByTrial.csv

**Description:**contains oral cavity volume measurements constructed from alpha hulls applied to the skeletal elements throughout each trial. Column names indicate the data category and rows correspond to frames, where all trials are concatenated by row. DoubleFullVolume is in cm3 and was calculated as 2x the alpha hull volume, as the alpha hull fills one half of the oral cavity from the body midline. Expansion rate is in cm3/sec and time is in milliseconds.

File: Kinematics_DataByTrial.csv

**Description:**contains rotations and translations of each skeletal element, calculated distances, individual, and key trial metadata for each trial. Column names indicate the data category and rows correspond to frames, where all trials are concatenated by row. Skeletal rotation columns are named by the measured (distal) element abbreviated (LJ = lower jaw, etc.) rel the proximal element abbreviated, followed by a period and two letters indicated whether the data are a rotation (r) or a translation (t) and about which axis. For example, CHrelNC.rz is the z-axis ceratohyal rotation about an ACS paired to the neurocranium. Locator positions are included for points that were used to calculate Euclidean distanced (neurocranial tip NCtip, lower jaw tip LJtip, etc. Rotations are in degrees. Time is in milliseconds, and each trial is centered to time zero at peak gape. Prey motion and other linear distances (Gape, SHmed, and SHlat) are measured in millimeters, with prey velocities in mm/s and accelerations in mm/s2, respectively. Gape and sternohyoid lengths (SHmed at the medial-most edge of the muscle, SHlat at the lateral-most edge) were zeroed by subtracting the average motion over 25 frames (50ms) at the start of the digitized trial, a period of time when the fish was not performing any cranial motions. Therefore, any changes are from a zero length of the position at rest. Rotational velocities are names as bone abbreviation z Vel and are measured in degrees per second. Trials are exported by date, time, and event in the format YYYY-MM-DD_HR-MN_Event#. For example, 2021-04-16_11-22_Evt09 was the 9th recording taken on April 16th, 2021 and exported from the internal camera memory at 11:22am local time. Food indicates the type of prey fed, with comet and large comet being separated by feeder fish that were greater or less than approximately 3cm.

File: StrikeMetadata.csv

**Description:**contains key trial metadata information, like fish ID, behavior, and food type. Column names indicate the type of metadata and rows correspond to trials. Trials are exported by date, time, and event in the format YYYY-MM-DD_HR-MN_Event#. For example, 2021-04-16_11-22_Evt09 was the 9th recording taken on April 16th, 2021 and exported from the internal camera memory at 11:22am local time. Associated CT scan OBJs and location of relevant processed files in the University of Chicago XMA Portal System are included, along with information about which frames are tracked in the uploaded XMAlab files. Also includes number of food items fed that day, where the FoodNumDay column indicates whether that strike was on the 1st, 2nd, etc. food item presented on that day of data collection. One fish (B6) occasionally spat out and re-consumed worms, so the data collected on 4-16-24 all indicate strikes on FoodNumDay 2 because the worm was not consumed between strikes. Other metadata (such as details of animal behavior, voltage and amperage of x-rays generated) are listed in the University of Chicago XMA Portal System for each trial and denoted in the TaskParameters .csv files in the Portal under Metadata folders that correspond to each day.

Code/software

R code packages for widely used XROMM analysis are referenced in the paper and accessible on Zenodo (https://doi.org/10.5281/zenodo.7734803). Custom code for calculating rotation onsets is available as a supplemental file “Calculate_zRotOnsets.R”

R version 4.3.1

Packages: dplyr and zoo

Access information

Other publicly accessible locations of the data:

xromm.rcc.uchicago.edu (https://xromm.rcc.uchicago.edu/larequest.php?request=explorePublicMetadataStudy&StudyID=7&instit=portal_base)

Version changes

March 21, 2025: Addition of custom code "Calculate_zRotOnsets.R" to Dryad repository.