ference curve for the small Pacific whiting closely 

 follows the selection curve in which euphausiid-sized 

 prey dominated. 



The selection curve for the middle-sized group of 

 Pacific whiting, 350-549 mm, also peaks sharply but 

 with long, trailing ends indicating that euphausiids 

 dominated the diet in numbers although small num- 

 bers of prey items, both smaller and larger than 

 euphausiids, were eaten (Fig. 6). The preference curve 

 is bimodal with the largest mode corresponding to a 

 predator-prey size ratio of about 100:1, while the 

 second mode reflects a continuing preference for 

 euphausiid-sized items. 



The largest size group of Pacific whiting, 550+ mm, 

 have an actual prey-size curve which depicts the 

 numerical dominance of euphausiids in their food by 

 the sharp peak at In (ivjii') value of about 10.5 (Fig. 

 7). There is also a pronounced hump in the left tail of 

 the curve in the region of a predator-prey weight ratio 

 corresponding to a large Pacific whiting predator and a 

 Pacific herring-sized prey. The preference curve shifts 

 completely away from the selection curve for these 

 Pacific whiting. The mode for euphausiid prey is 

 dampened almost completely and the most prevalent 

 predator-prey size ratio is about 130:1, equivalent to 

 the ratio of a large Pacific whiting predator to a 

 Pacific herring-sized prey. 



Thus, it appears that the diets of Pacific whiting 

 <200 mm long reflect a preference for euphausiid- 

 sized prey. Although the 350-549 mm size group of 

 Pacific whiting shows a dominant prey-size preference 

 of 100:1, the calculated preference curve deviates 

 from the predicted normal shape due to a second mode 

 in the region corresponding to euphausiid-sized prey. 

 Deviations from normality can be caused when the 

 abundance of a naturally occurring prey item is not 

 inversely proportional to its weight as assumed or 

 when the prey occur in such dense patches that the 

 predator consumes more than one prey item at a time 

 (Ursin 1973). The latter case is a likely description of 

 Pacific whiting predation on euphausiids, though this 

 would have to be verified through direct observation. 

 The largest size group of Pacific whiting shows a pre- 

 ference for Pacific herring-sized prey with a median 

 predator-prey size ratio of about 130:1. This is 

 similar to Ursin's (1973) calculation of an average 

 predator-prey size ratio of 160:1 for Atlantic cod, 

 Gadus morhua. 



The prey-size preference of Pacific whiting is 

 reasonably described by Ursin's model. The major 

 parameters which define the shape of the prey-size 

 preference curve are the mean (x) and variance (s 2 ) 

 of the frequency distribution of predator-prey size 

 scores (Table 3) when equal numbers of prey sizes 



are offered to the predator. These parameters are Jc = 

 7.16,s 2 = 0.42 for < 200 mm fish; x = 6.48, s 2 = 7.18 

 for 350-549 mm fish; andx = 4.88, s 2 = 2. 40 for 550 + 

 mm fish. 



Diel Feeding Pattern 



A total of 258 Pacific whiting stomachs (7 of which 

 were empty) were taken at 7 different times of day 

 during a 15-h period at a location off the Washington 

 coast in July 1967. To detect any discontinuity in 

 feeding during this period, a one-way analysis of 

 covariance was performed using the model 



y = fi + a t + 6x„ 



wherey is the weight of the stomach contents and x is 

 the Pacific whiting weight. If the stomachs are 

 grouped by time, i, then the test of among-time varia- 

 tion in the weight of stomach contents after adjust- 

 ment for Pacific whiting weight is the test of the 

 equality of the intercepts, a„ given a common slope, b 

 (Jenkins and Green 1977). The F-ratio for this testis 

 the among-group variance estimate divided by the 

 within- group variance estimate. 



Figure 8 plots mean stomach content weight as a 

 percentage of Pacific whiting weight for samples 

 taken during the 15-h period in July 1967. Stomach 

 content weight per fish weight is highest at 1800 h, 

 with a value of 2.5%, and slightly increases between 

 0100 and 0300 h and between 0400 and 0900 h. De- 

 spite the great variability among all time periods, the 

 analysis of covariance of the data rejects the null 

 hypothesis of no difference between group means of 

 stomach content weight at the 0.01 level of 

 significance [F(6,243 df) = 7.83]. Therefore, feeding 

 by Pacific whiting was discontinuous during the sam- 

 pling period. 



Stomach fullness was highest in early evening ( 1 800 

 P.S.T.), with some increases after midnight (0300 

 P.S.T.) and morning (0900 P.S.T.). These fullness 

 peaks coincide with the times of euphausiid, 

 Thysanoessa spinifera, concentration in the same 

 portion of the water column where the Pacific whiting 

 are concentrated— nearbottom at evening and morn- 

 ing and nearsurface after midnight (Alverson and 

 Larkins 1969; Alton and Blackburn 1972). 



Hickling (1927) noted that stomachs of European 

 hake, Merluccius merluccius, were fullest at midnight. 

 He also suggested that European hake migrate ver- 

 tically in search of euphausiids. Silver hake, Merluc- 

 cius bilinearis, of the northwest Atlantic also feed 

 nocturnally, starting their feeding activity after dusk 

 and actively feeding until midnight (Bowman and 

 Bowman 1980). Thus, fish of the genus Merluccius 



634 



