FISHERY BULLETIN: VOL. 83. NO. 3 



sample and the remaining ovarian material were 

 blotted on absorbent paper and weighed to the 

 nearest milligram. The subsample was then placed in 

 modified Gilson's fluid (Bagenal and Braum 1971), 

 teased apart and vigorously shaken to separate ova, 

 and stored for several weeks. All eggs in a subsample 

 were counted and classified as belonging to either 

 the recruitment pool or the mature batch of eggs. 

 Batch and recruitment fecundities were then esti- 

 mated for each female by multiplying the number of 

 eggs in the subsample by an expansion factor (total 

 ovarian weight/subsample weight). The mean 

 diameter of the largest size class of eggs was also 

 determined by measuring a random sample of about 

 20 mature eggs from each female. Shrinkage of eggs 

 during preservation appeared to be minimal, so no 

 correction for shrinkage was made. Fecundity was 

 estimated for females collected on four occasions, 

 from just before the beginning of the spawning 

 season to its end. These dates were 6 May, 6 June, 22 

 June, and a pooled sample captured over the period 

 6-13 July. 



To demonstrate whether estimates of fecundity 

 were dependent on the location of the ovarian sub- 

 sample, I compared these among replicate subsam- 

 ples taken from the anterior and posterior sections of 

 the right and left ovaries of six different females (i.e., 

 four subsamples per female). Two-way ANOVA in- 

 dicated that estimated fecundity was independent of 

 subsample location (F > 0.10) and the coefficient of 

 variation was low (CV = 4.2%). The ratio of the 

 number of batch eggs to recruitment eggs was also 

 independent of subsample location (P > 0.5). More- 

 over, I also directly counted the total number of 

 batch eggs in four of the females used in the above 

 analysis; in each case estimated batch fecundity was 

 within 10% of the true value. 



Field Observations of 

 Spawning Frequency 



The frequency of spawning in a natural population 

 of Menidia menidia was inferred from daily, direct 

 field observations of mating. The Atlantic silverside 

 spawns in large groups of fish that broadcast milt 

 and eggs amongst vegetation in the upper intertidal 

 zone of salt marshes within 1 or 2 h of the daylight 

 high tide (Middaugh 1981; Middaugh et al. 1981; 

 Conover and Kynard 1984). At such times, spawning 

 can be easily observed. My observations were con- 

 ducted at a major spawning site in Salem Harbor, 

 MA. Daily assessments of spawning intensity were 

 conducted throughout the spawning season by count- 

 ing the number of aggregations of spawning fish 



sighted during high tide. Methodological details are 

 provided in Conover and Kynard (1984). 



Laboratory Observations of 

 Spawning Frequency and Egg Production 



Spawning frequency and egg production were also 

 assessed by confining fish in laboratory tanks. A 

 summary of the experimental procedure, described 

 fully in Conover and Kynard (1984), follows. A large 

 group of M. menidia were captured at Salem Harbor 

 on 5 May 1979 and transported to the University of 

 Massachusetts marine laboratory at Gloucester, MA. 

 One female and two male fish were placed into each 

 of the four 74 L tanks at room temperature on a 

 natural photoperiod. Four males and four females 

 were also placed into each of two circular plastic 

 pools (diameter 1.5 m, depth 0.3 m). These pools 

 were located outdoors. All fish were fed fresh, chop- 

 ped seaworms (Nereis) and amphipods in excess of 

 daily consumption. Spawning substrates consisted of 

 a small tuft of synthetic aquarium filter floss, an- 

 chored to the bottom of each tank or pool. The floss 

 was checked several times daily for the presence of 

 eggs. When eggs were discovered, the floss was 

 replaced. All eggs were preserved and enumerated 

 later. Eggs from each female were usually deposited 

 in a distinct clump on the floss, providing a means for 

 determining the number of females that had spawn- 

 ed in the previous interval. The experiment was 

 allowed to continue until spawning ceased (27 July). 

 Length and weight of spawners was measured at the 

 conclusion of testing. 



RESULTS AND DISCUSSION 



Gonadal Analyses 



In late fall, just prior to the offshore winter migra- 

 tion (Conover and Murawski 1982), ovaries repre- 

 sented about 1% of total body weight (Fig. 2) and 

 contained only small (< 0.1 mm), clear, transparent 

 eggs. Upon return of fish to the shore zone the 

 following April, the ovarian GSI was about 4% and 

 many opaque, white, immature ova (< 0.5 mm) were 

 present. Of the 25 females captured on 6 May, 92% 

 contained numerous immature ova and a clearly 

 definable batch of maturing ova. The first female 

 carrying ripe egges was collected on 12 May. The 

 GSI peaked in both sexes on 25 May and declined 

 thereafter until the end of July. The first fish in spent 

 condition (no maturing egg class, recruitment eggs, 

 if present, degenerating, GSI < 5%) was captured on 

 22 June. The proportion of spent fish was 23% on 



334 



