Collection of adults and spawning:  Adults of each species were collected from intertidal or shallow subtidal zones from various sites in Los Angeles County and transported to California State University Long Beach, where they were maintained in recirculating seawater tanks at 16 °C until their use in experiments. Experiments were carried out on one species at a time, depending on reproductive seasonality for that species.

Spawning was induced using standard methods (e.g., M. Strathmann, 1987). The echinoids Dendraster excentricus, Lytechnius pictus, and Strongylocentrotus purpuratus were induced to spawn via injection of 0.2-1.0 mL (depending on adult size) 0.53 M KCL into the perivisceral coelom. The asteroids Patiria miniata and Astropecten armatus were induced to spawn by injection of 1-3 mL 100 µM 1-methyladenine. The holothuroid Apostichopus parvimensis was injected with 3 mL of 200 µM NGLWY-amide (Kato et al., 2009). The ophiuroid Ophiothrix spiculata was exposed to 4 °C water in the dark for 15 minutes, then to room temperature water and sunlight for 15 minutes; this treatment was repeated for up to two hours (Selvakumaraswamy & Byrne, 2000). For all species, adults were each induced to spawn in their own isolated containers, allowing us to control subsequent fertilizations.

For each species, we prepared four full-sibling families, each with a unique pair of parents. Gametes were collected from four parents of each sex. Eggs of each female were rinsed and resuspended in FSW. Two to four concentrated drops of sperm from each male were diluted in 50 mL FSW. The diluted sperm of a single male were then added to the rinsed eggs of a single female a few drops at a time until fertilization approximated 90%. Fertilized eggs were then diluted with FSW and allowed to develop undisturbed for 24-30 h at 16 °C. By that time embryos had reached the swimming blastula or gastrula stages (depending on species) and were concentrated at the water’s surface. We decanted the embryos of each unique family into a clean beaker, stirred them well, then estimated the concentration of embryos in each of these stock suspensions (each representing a unique family) by averaging the number of larvae in five 0.5 mL subsamples. For each species, we sought to expose larvae of four unique families to low and high food treatments in a paired (by family) design. Thus, for each of the four unique families, we prepared two beakers, each containing 1 L of FSW. To each of these beakers we added the appropriate volume of stock suspension to deliver an estimated 250 embryos from that family.

Culturing: One beaker per family was fed the low food ration (1000 cells mL−1 Rhodomonas lens) and the other was fed the high food ration (6000 cells mL−1 R. lens). This resulted in eight beakers per species, with each low-fed beaker paired with a high-fed beaker of the same family. Rhodomonas lens were isolated from their growth medium via centrifugation, resuspended in FSW, and counted using a BD Accuri C6 flow cytometer. Cultures were maintained in the 16 °C environmental chamber and continuously stirred by a paddle system (Strathmann, 1987). Daily water changes began on the third day post fertilization (dpf), allowing larvae to develop without disturbance for a day while still in or just completing the pre-feeding period. To change the water, cultures were filtered through a 60 µm sieve to capture larvae. The sieve was submerged in shallow water while filtering so larvae were not exposed to air. Larvae were then gently rinsed from the filter with fresh FSW back into their cleaned beakers and fed.

Sampling: We imaged larvae at three developmental timepoints: just prior to the onset of their ability to feed (“pre-feeding”), and at two timepoints after the onset of feeding (“early” and “mid”). The timing of pre-feeding, early, and mid-development measurements varied among each of the species depending on their rate of development. For pre-feeding measurements, we aimed to sample larvae just before the onset of feeding in the culture. These timing of the onset of feeding was determined from prior observations (e.g. Pernet et al., 2017), or, in cases where prior observations of development at similar temperatures were not available, by rearing a preliminary culture in the same conditions as the main experiment, checking larvae in that culture for evidence of feeding (algal cells in the stomach) at least three times daily.For the later timepoints, we aimed to measure larvae before there was visible evidence of a rudiment in echinoids and asteroids. For most species, 7 dpf was the default day for early-development images, and 14 dpf was the default day for mid-development measurements. These measurements were taken earlier in echinoids due to their rapid development, the most extreme being 3 dpf and 7 dpf for D. excentricus, followed by 5 dpf and 10 dpf for L. pictus, and 6 dpf and 12 dpf for S. purpuratus. Despite these precautions we observed that many D. excentricus had begun rudiment development by 7 dpf, and some S. purpuratus were at the very earliest stages of rudiment invagination (as described by Heyland & Hodin, 2014) at 12 dpf.

At these timepoints, we measured ciliated band length as described above. Body length was also estimated from two landmarks. Three additional parameters – mouth width, stomach length, and stomach width – were measured in two dimensions using a single image from the stack in which the structure of interest was best in focus. The geometric mean of stomach length and width was used to represent stomach size (Strathmann, 2022). Additionally, in bipinnaria and auricularia larvae, the ciliated band was traced from a single image as depicted in Caballes et al. (2017) to evaluate it as a proxy for the direct measurement of the ciliated band length.