**Initiation of feeding**

Impacts of water treatments on development of larval particle feeding mechanisms were determined by measuring the proportion of mussel larvae from each treatment that ingested fluorescent beads at 44 h post-fertilization (initiation of feeding, IF). Preliminary experiments demonstrated that at 44 h after fertilization >50% of *M. californianus* larvae began feeding when reared at ambient PCO2 (~380 ppm) and 18C.

We expanded upon previous IF findings and determine the length of the delay of the onset of feeding and how this delay affected the modeled growth of larvae to 260 um in shell length, the size at which larvae typically develop into pediveligers. We quantified the delay by first determining the relationship between the proportion of larvae feeding under optimal conditions and time since fertilization. This relationship was best described by the following three parameter logistic equations:

% Feeding= 94.1/((1+Exp (-0.74 × (h-45.1))))

where *h* is the hour post-fertilization. The logistic equation was then rearranged and linearized, enabling us to estimate the functional age of larvae feeding in each ocean acidification (OA) treatment, by comparison with the proportion of larvae feeding under normal conditions.

**Particle processing**

To assess the effects of OA on particle processing, 48 h old larvae from each treatment were stocked in nine 25 ml VOA vials (10 larvae/ml) containing the same water treatment in which they developed from fertilized eggs. After an acclimation period of one hour, larvae were then exposed to 2 um Fluorescbrite Polychromatic (Polysciences Inc., Warrington, PA) yellow (Y) beads (excitation maxima of 441 nm and emission maxima at 485 nm) at a concentration of 20 beads/ul and allowed to feed on these beads for one hour. A second and equal dose of 2 um red (R) beads (excitation maxima of 491 nm and 512 nm and emission maxima at 554 nm) were added to the vials at a concentration of 20 beads/ul following the hour-long exposure to Y beads. Triplicate vials were assigned to one of three exposure groups (10, 30, and 50 min) after red beads were added to the vials. To terminate feeding activity at the prescribed exposure time and preserve larvae for later analysis, 40 ul (0.2% v/v) of 10% buffered formalin (pH = 8.1-8.2) were added to vials. Later, larvae were crushed under a cover slip to flatten gut contents and allow better enumeration of all ingested beads in larvae under an epifluorescent microscope (objective 20x; Leica DM 1000). Larval sample sizes consisted of greater than or equal to 20 larvae per replicate vial per treatment.

**Gut fullness **

Gut fullness was defined as the mean total number of ingested beads (Y+R beads) per larva over 10, 30, and 50 min sampling periods.

**Ingestion rate**

Ingestion rates were estimated by determining the uptake of R beads after the first 10 min of exposure to this bead type. We then doubled the number of ingested beads as larvae were found to consume R and Y beads at equal rates in preliminary experiments.

**Standardizing particle processing for shell-length effects**

We examined the relationship between larval shell length (SL), gut fullness, and ingestion rate from a subset of treatments spanning the range of experimental omega-aragonite categories (greater than or equal to 10 larvae from 10 different VOA vials). Shell lengths, defined as the longest axis parallel to the shell hinge, were obtained by photographing larvae under a light microscope (50x) and measuring shell lengths using Image-pro (v.7).

After finding a significant relationship between larval shell size and feeding metrics, we applied the following hyperbolic function from Waldbusser et al. (2015), which strongly predicted the shell lengths of these larvae from the omeaga-aragonite for the first 48 h of development, to estimate shell lengths of larvae for all treatments:

SL=(884.378 × OM_ar )/(1+7.691 × OM_ar )

Next, we divided gut fullness values and ingestion rates of each treatment by their shell length estimate using the above equation. We then reexamined the effects of carbonate chemistry parameters on these feeding metrics after accounting for shell length.

**Modeled effects of initial 48 h OA exposure on subsequent larval energy budgets, growth and development**

To estimate the effects of exposure to OA treatments during the first 48 h of larval development on subsequent energy budgets of M. californianus larvae, we first estimated energy contents of larvae in treatments at 44 h post-fertilization by applying the relationship between estimated total body energy content (E_SL; uJoules) and larval shell size reported for Mytilus edulis larvae by Sprung (1984a):

E_SL = 2.28 × (10^-7) × (SL^3.12)

where SL is the shell length (um). To our knowledge, this is the only known allometric relationship between energy content and shell length for any larval Mytilus species. Additionally, the relationship provided by Sprung (1984a) seemed appropriate to use here as M. edulis and M. californianus are similar with respect to egg size and, presumably, energy content (Strathmann 1987).

The total energy contents of larvae (ET) at 48 h post-fertilization (i.e. after a 4 h period during which larvae could be feeding on algal cells) were then estimated to evaluate the energy budgets of larvae among OA treatments. ET was estimated by accounting for ESL and other potential energy gains and losses using the following equation:

ET = ESL + ((I×AE)/h × (4-D)) - R

where I is the ingestion rate (uJoules/h), AE is the assimilation efficiency (%) that describes the conversion of ingested food to energy available to larvae, R is the energy loss due to respiration (uJoules) over the 4 h period, and D is the energy gain or loss (uJoules) to ET as a result of developmental advancement or delay in feeding activity. Physiological rate processes were converted to energy units under the following assumptions: 1) growth or change in ESL was negligible between initiation of feeding experiments (44 h) and when the sizes of larvae were measured at 48 h; 2) larval ingestion rates were constant over the period from 44 to 48 h and were equivalent to algal ingestion rates with an estimated algal cell energetic content of 0.61 uJoules/cell (Isochrysis galbana, Sprung 1984a), which was also the food source underpinning the relationship between energy content and shell length; 3) total time for potential gains from ingestion was 4 h +/- any advancement/delay in the initiation of feeding among early larvae; 4) assimilation efficiencies of larvae were 0.38 (Sprung 1982); 5) respiration rates were similar among larvae in all treatments except for those of larvae in the lowest pH treatment (based on Waldbusser et al. 2015); and 6) 1 nl O2 was equivalent to 20.1 uJoule of respired energy (Crisp 1971).

To estimate the impacts of OA during the first 48 h of development on subsequent larval growth, we modeled the developmental time of larvae to reach a shell length of 260 um, the approximate size of pediveliger mussel larvae under normal conditions (Sprung 1984a) and a proxy for larval competency for settlement and metamorphosis; however, we note that larval competency is frequently correlated with but not necessarily dependent on larval size (Coon et al. 1990, Pechenik et al. 1996). To make these extrapolations, we first estimated energy content gains of larvae in 24 h intervals (delta E) using the following equation:

delta E= ESL + (I × AE × NGE)

where ESL is the estimated energy content of the larvae based on their shell length at 48 post-fertilization, I is the ingestion rate (uJoules/h), AE is the assimilation efficiency (%), NGE is the net growth efficiency (%). ESL and I in this study were assumed constant between 48 and 72 h post-fertilization. After this initial 24 h period (i.e. 48-72 h post-fertilization), modeled gains in larval energy content were added to those of larvae of each treatment and larval shell lengths were adjusted by rearranging equation 2. From 72 h post-fertilization onward, we modified larval ingestion rates and larval respiration rates in accordance with allometric equations for M. edulis as described by Sprung (1984b, c). This model explicitly tests how an acute initial 48 h OA exposure could significantly alter the subsequent duration and energy budgets of mussel larvae. Assimilation efficiencies were 0.38, 0.29, and 0.27 for larvae < 200 um, between 200-250 um, and > 250 um, respectively (Bayne 1983). NGE were also adjusted for changes in larval size and estimated at 0.78, 0.67, 0.65 for larva < 200 um, between 200-250 um, and > 250 um (Bayne 1983). The impacts of differences in initiation of feeding, initial feeding rates and shell size after 48 h of development on subsequent larval growth to 260 um were estimated by non-linear multiple regression analysis.