Direct effect of dissolved oxygen on predation by juvenile striped bass.--As with sea nettle experiments, striped bass predation experiments were set up as randomized complete blocks (by date of replicate) with paired predator and no-predator controls at each dissolved oxygen concentration.
As with sea nettle experiments, the order of the addition of animals to mesocosms was zooplankton, then larvae, and then, where appropriate, predators.
Because there was no effect of dissolved oxygen concentrations [is greater than or equal to] 2.5 mg/L on the recovery of larval prey in either the striped bass experiments testing the direct effect of dissolved oxygen, or in the sea nettle mesocosm experiments, we interpret larval recovery in these interaction experiments to be directly proportional to the number of larvae not eaten.
Measured dissolved oxygen, temperature, sea nettle bell diameter, sea nettle volume, and proportion recovery rate of prey in no-predator controls in small-scale predation experiments.
Sea nettle predation on larvae.--The proportion of naked goby larvae captured by sea nettles in 1-[m.sup.3] mesocosms increased significantly at reduced dissolved oxygen concentrations from a low of 0.26 [+ or -] 0.03 in air-saturated controls to a high of 0.87 [+ or -] 0.02 in mesocosms maintained at 1.5 mg/L (Fig.
MANOVA results for sea nettle mesocosm experiments.
Neither average bell diameter, nor total sea nettle volume in mesocosms significantly affected the proportion of large or small naked goby larvae eaten when these measures of sea nettle size were added to the randomized block MANOVA as covariates (Wilk's [Lambda]: number df = 2, density df = 9, P = 0.536 and 0.455 for volume and bell diameter, respectively).
The arcsine-transformed proportion of larvae eaten was significantly affected by both dissolved oxygen treatment and sea nettle diameter (ANCOVA: F = 66.4, P = 0.0001, df = 1, and F = 6.00, P = 0.04, df = 1 for dissolved oxygen and sea nettle diameter, respectively), but not by sea nettle volume (P = 0.27 for volume covariate in ANCOVA).