The synergic relationship between physiology, ecology and evolutionary process makes the body size distribution (BSD) an essential component of the community ecology. Body size is highly susceptible to environmental change, and extreme upheavals, such as during a mass extinction event, could exert drastic changes on a taxon’s BSD. It has been hypothesized that the Late Triassic mass extinction event (LTE) was triggered by intense global warming, linked to massive volcanic activity associated with the Central Atlantic Magmatic Province. We test the effects of the LTE on the BSD of fossil bivalve assemblages from three study sites spanning the Triassic/Jurassic boundary in the UK. Our results show that the effects of the LTE were rapid and synchronous across sites, and the BSDs of the bivalves record drastic changes associated with species turnover. No phylogenetic signal of size selectivity was recorded, although semi-infaunal species were apparently most susceptible to change. Each size class had the same likelihood of extinction during the LTE, which resulted in a platykurtic BSD with negative skew.  The immediate post-extinction assemblage exhibits a leptokurtic BSD although with negatively skewed, where surviving species and newly appearing small-sized colonizers exhibit body sizes near the modal size. Recovery was relatively rapid (~100kyr), and larger bivalves began to appear during the Pre-Planorbis Zone, despite recurrent dysoxic/anoxic conditions. This study demonstrates how a mass extinction acts across the size spectrum in bivalves and shows how BSDs emerge from evolutionary and ecological processes.

Morphometric measurements of bivalves were collected from the base of the Westbury Formation (Rhaetian) to the Angulata Zone (Hettangian) in St. Audrie’s Bay; from the base of the Langport Member to the Angulata Zone in Pinhay Bay; and from the Westbury Formation to the Liasicus Zone (Hettangian) in Larne (Supplementary Figure 1 and supplemental files). One sample of 1.5 ± 0.2 kg was collected from each major lithology (limestone and mudstone) on average every 1.98 ± 0.5m. In addition, to increase the sample size of each stratigraphic horizon, individual measurements were made in the field from bedding plane exposures using a tape measure (standard error ± 0.5 mm).

Samples from the different lithologies were processed in different ways. Limestone rock samples of 1.5 ± 0.2 kg (with dimensions of ca. 20cm length × 20cm width ×15cm height) were cut perpendicular to the bedding plane surface, generating slabs approximately 8cm wide (see Supplementary Figure 2A). Using a chisel-tipped hammer, each slab (20×15×8 cm) was then splitted parallel to the bedding plane, which generated sampling plates of approximately 2×20×8cm with specimens exposed on the surfaces of each slab for potential measurement (see Supplementary Figure 2A). This technique minimized damage to the specimens in order to preserve entire individuals for measurement.

In contrast, mudstone samples were broken into ~49 cm² chips, and water was used to soak the rock and separate the fossils without damaging them. All the bivalves were classified from published information (Hodges 2000, Lord and Davis 2010, Swift and Martill 1999). In all cases, measurements of the length (maximum distance on the anterior–posterior axis) and height (maximum distance on the dorsal-ventral axis) were recorded only from complete, well-preserved specimens (Supplementary Figure 2B-C). All morphometric variables were measured using electronic calipers (Standard error ± 0.01 mm).

Data Analysis: The body size of each specimen was calculated as the geometric mean (i.e., the square root of the product of length and width) and log₂ transformed to reduce the heteroscedasticity and to facilitate graphic representation (Piazza et al. 2019).

Supplementary information contains:
 Dataset.Body.Size.Opazo&Twitchett2021.csv_Paleobiology contains one spreadsheet with all body size measurements used in this study. This file is made up of 17 columns which indicates: 


• Formation = Corresponds to stratigraphic units (and biozones) considered in this study: 
WF: Westbury Formation/ CM: Cotham Member/ LM: Langport Member/ PPZ=Pre-Planorbis Zone/ PZ=Planorbis Zone/ LZ: Liasicus Zone/ AZ: Angulata Zone
• 
Locality = NI: Larne, Northern Ireland/PB: Pinhay Bay, Devon/AB: St. Audrie's Bay (all from United Kingdom)
• Sample = Sample Code
• Heights (m) = Stratigraphic height
• Tier = Infaunal/Surficial/Semi-infaunal
• Taxonomy=Order/Family/Genus/Species


Ecological categories: Colonizers/Extinct/Survivors
Length (mm) = Maximum distance on the anterior–posterior axis

• Height (mm) = Maximum distance on the dorsal-ventral axis
  • BS (mm) = Body size expressed as geometric mean (i.e., the square root of the product of length and width)
• log2 (BS) = log2 of the geometric mean
• Residuals =  Grand bivalves mean (log2) minus the body size (log2) of each specimen
• Lithology = Mudstone/Limestone

Opazo&Twitchett2021.Supplementary.text.pdf_Paleobiology: This file contains information regarding paleoenvironment, extinction event and the Triassic/Jurassic Boundary, and geological setting of the Triassic/Jurassic Boundary in the United Kingdom. Additionally, this file contains the R code to generate a null model completely random.
    1.    Opazo&Twitchett2021.Supplementary.figures.pdf_Paleobiology
    2.    Opazo&Twitchett2021.Supplementary.tables.pdf_Paleobiology