The following sections contain methodology excerpts from Quinlain et al. (2019) relevant to this dataset.

Crustose Coraline Algae Collection and Identification
Both Hydrolithon reinboldii and Porolithon onkodes were collected from Patch Reef 42 (21.4785˚, -157.8281˚) in Kāne'ohe Bay, O'ahu, Hawai'i on 4 May 2017. Porolithon onkodes is a common CCA species in the Pacific Ocean that is often used for larval settlement experiments with coral species in Australia (Heyward and Negri 1999). It is typically found in high light and high flow environments, such as at the top of the patch reefs in Kāne‘ohe Bay. This species is characterized by its smooth surface texture, and diagnostic depressions of trichosite fields. While there is a recent paper showing that this species is a species complex globally (Gabrielson et al., 2018), we retain the use of the name P. onkodes here to be consistent with the published taxonomic monograph for CCA in Hawaiʻi (Adey et al., 1982). Hydrolithon reinboldii is also a common CCA species that is found throughout the Pacific Ocean. It is known to facilitate coral larval settlement (Harrington et al., 2004). This species often lives cryptically in cracks in the reef or on the bottom of small pieces of calcium carbonate rubble. It is characterized by slightly raised hemispherical single pore conceptacles (400-600 µm in diameter), and a patchy surface texture referred to as tessellate (Adey et al., 1982).

Fragments of both species of CCA were trimmed using bone cutters to ensure only a single plant was on each fragment. Each fragment still retained bare calcium carbonate along with the individual species of CCA. To control for the bare calcium carbonate, encrusted fragments of calcium carbonate were similarly trimmed to remove any small CCA plants and epiphytes leaving only the calcium carbonate rubble and endophytes. After fragmentation the specimens were haphazardly placed into six containers and randomized within a 1300 L flow through seawater bath to maintain all treatments at a stable temperature, which was the same as those found in Kāne‘ohe Bay. As there are currently no studies on the effect of fragmentation on exudate production we allowed the fragmented algae to recover for five days before starting the exudation experiment. Flow through seawater baths were covered by shade cloth to reduce natural irradiance to levels similar to those found at depth in Kāne‘ohe Bay where both species are naturally found. Both species were exposed to the same light levels as to not bias by variation of abiotic parameters.

Incubations and sample collection
Twenty-four 250 mL glass beakers were washed with 10% volumetric HCl, rinsed with milliq-water and air-dried. At 07:30 on 9 May, 3 L of seawater (sand filtered and collected from the Hawai‘i Institute of Marine Biology flow-through seawater system in Kāne'ohe Bay) was vacuum pre-filtered through 0.2µm polyethersulfone filters (47 mm; Sterlitech) in a 500 mL polysulfone graduated filter holder. Before water was aliquoted into the beakers, samples for fluorescent DOM (fDOM), dissolved organic carbon (DOC), and flow cytometry (FCM) were collected from the 500 mL polysulfone graduated filter holder. Each beaker was filled and randomized within a 1300L flow through seawater bath to maintain stable temperature between the treatments. Each organism treatment beaker (water control, calcium carbonate control, Hydrolithon reinboldii, or Porolithon onkodes) was filled with seawater (filtered or unfiltered) and replicated (n = 3) for a total of 24 beakers (4 organismal treatments * 2 water treatments * 3 replicates). Filtered and unfiltered treatments were designed to capture differences in sloughing behavior between species. A Multiple trimmed fragments of each organism were placed within their respective beakers so that the total surface area within each replicate beaker was standardized to 20-30 square cm (25.57 ± 4.13 cm2). The incubation began at 9:00 and was halted at 17:00 to maintain only exudates produced during the daylight hours. Surface area was digitally determined at the end of the experiment by analyzing images to scale with image-J (Schindelin, Arganda-Carreras, & Frise et al, 2012).

DOM samples were collected at the beginning of the experiment before aliquoting the water at 9:00 and from each beaker at 17:00. DOM samples were immediately filtered through a 0.2 µm polyethersulfone filter (47 mm; Sterlitech) in a 500 mL polysulfone graduated filter holder. Filtrate was poured directly from polysulfone graduated filter holder into its respective sample vial, Filtrate for fDOM samples were collected in acid washed, combusted, triple sample-rinsed amber borosilicate vials with Teflon septa caps and stored dark at 4˚C until analysis for fDOM within 24 hours. DOC was collected in acid washed, combusted, triple sample rinsed clear borosilicate vials with Teflon septa caps and measured as non-purgeable organic carbon via acidification, sparging and high temperature platinum catalytic oxidation on a Shimadzu TOC-V at the UCSB DOM Analytical Lab following the methods outlined by Carlson et al. (2010). Samples for flow cytometry were collected by pipet (1 ml amended to a final concentration of 0.5% paraformaldehyde, mixed by inversion, snap frozen -80ºC) at 9:00, 13:00, and 17:00.

Sample analysis
Flow Cytometry: Flow cytometry was used to measure total nucleic acid-stained cell concentrations. Samples were thawed and 200 µL were aliquoted into u-bottomed 96-well autosampler plates and stained with 2 µL of 100X SYBR Green I stain (final concentration of 0.5X). Samples were analyzed on an Attune Acoustic Focusing Cytometer with Autosampler Attachment (Life Technologies, Eugene, OR, USA). Samples were run at a flow rate of 100 µL min-1 on standard sensitivity; 150 μL of sample was aspirated, 75 μL was counted and data was collected only from the last 50 μL (event rates were empirically determined to be steady only after 25 μL of continuous sample injection per Nelson et al., 2015).

Fluorescence spectroscopy: Samples for fluorescence spectroscopy were measured using an Horiba Aqualog scanning fluorometer following the methods of Nelson et al. (2015), including scan time and resolution, spectral data processing, inner filter correction, Raman unit standardization, blank subtraction and PARAFAC modeling (Stedmon and Bro 2008; Lawaetz and Stedmon 2009; Kothawala et al. 2013). Scans were processed using a Matlab (v2007b) script written and specified by Nelson et al. (2015) and Quinlan et al., (2018; most recent version available at DOI: 10.5281/zenodo/3479841), modified to additionally capture the peak present at Excitation 240 nm and Emission 300 nm (phenylalanine-like: Lakowicz 2010). Six modeled components were validated using split half validation and outlier analysis (Quinlan et. al., 2018). All PARAFAC components had similar excitation-emission maxima and strong covariation among samples with previously identified fluorophores (Quinlan, et. al., 2018); for subsequent analyses we examined established fluorescence maxima from the literature (Coble 1996; Stedmon et al. 2003; Lakowicz 2010).