544 



Fishery Bulletin 105(4) 



between cod (GM and GOM combined) and plaice in the 

 freezing treatment (P<0.001). 



For four haddock samples in the freezing treatment, 

 mean oocyte diameter was measured immediately af- 

 ter the sample was freshly thawed, and then again 

 after 48 hours of refrigeration. Mean oocyte diameter 

 of the samples decreased by 3-10% after refrigeration, 

 although the average decrease of 4.4% was not signifi- 

 cantly different from zero (P=0.11), perhaps because of 

 the small sample size. 



Discussion 



Of the preservatives tested in this study, the results 

 with formalin were most similar among individuals and 

 species, and the results from freezing were the least 



100 —I 



80 — 



g 



a 60 



Atlantic cod, combined 

 Atlantic cod, Georges Bank 

 Atlantic cod. Gulf of Maine 

 Haddock, Georges Bank 

 American plaice, Gulf of Maine 



40 



20 — 



-40 



Formalin 



Figure 1 



Percent difference in ovarian material weight C^r 4,y,) for each species 

 (Atlantic cod [Gadus morhua]), haddock [Melanogrammus aegle- 

 finus], and American plaice [Hippoglossoides platessoides]) and 

 region, grouped by treatment. For Atlantic cod, observations are 

 displayed together in the combined group and are then displayed 

 separately by region (i.e., Georges Bank or the Gulf of Maine). The 

 horizontal line at zero on the vertical axis indicates values where 

 the percent difference in weight between preserved and fresh ovar- 

 ian material was equal to zero. Circles indicate outliers. Asterisks 

 indicate that mean '/f^w, was significantly different from zero at 

 the appropriate Bonferroni adjusted a-value. Box length is equal 

 to the interquartile range which contains the central 50% of the 

 observations (25% above and below the median). Whiskers extend 

 1.5 box-lengths above and below the limits of the box. 



consistent among individuals within and among species. 

 In terms of image quality, samples preserved in formalin 

 were clearest and contained the least debris. In weight 

 comparisons between cod and plaice, the subformalin 

 treatment was the only one where no significant differ- 

 ence between the two species was found. Of the tested 

 preservatives, the best option for a standard preservative 

 was formalin, especially when samples might be used for 

 histology, where postovulatory follicles and fine cellular 

 structures must be preserved. 



When realized fecundity is estimated from total po- 

 tential fecundity, estimates may be biased if samples 

 are collected too late in the season and spawning has 

 already begun, or if atresia is likely to occur between 

 sampling and the time of spawning. Both of these bi- 

 ases are quantified through histological analysis. In 

 species or stocks where the developing stage is difficult 

 to identify or where high rates of atresia 

 are expected, histological analysis is neces- 

 sary to assure that estimates of realized 

 fecundity are accurate. Although Gilson's 

 solution and ethanol may be used for fecun- 

 dity research and may be superior in certain 

 situations for particular species, they do not 

 preserve tissue well enough for histological 

 analysis, and thus they are undesirable in 

 many situations. 



Formalin is a common preservative for 

 ovarian material and is widely used to fix 

 and preserve animal tissue. Although our 

 results show differences between formalin 

 samples and Gilson's and frozen samples 

 that contrast with results of studies on other 

 species, the effects of formalin preservation 

 were very consistent among the species we 

 studied. Formalin preservation resulted in 

 an increase in sample weight in all species, 

 and when compared between species these 

 changes were similar. 



The differences between the formalin and 

 split-formalin treatments for all species (Fig. 

 2), although significant for GB and GOM cod 

 combined, GOM cod, and plaice, were quite 

 small and therefore indicated that formalin 

 preserves oocyte size similarly whether ovar- 

 ian material is in large membrane-bound 

 lobes or in 1.5-mL subsamples. However, the 

 large and significant difference in %A-^, of 

 samples of GOM cod ovarian material be- 

 tween the lobe-formalin and subformalin 

 treatments indicates a conflict because oo- 

 cyte diameter and ovarian material weight 

 are inherently related. If ovarian material 

 weight increases much more in one treat- 

 ment than another, oocyte size should as 

 well. The reason for this disagreement may 

 lie in the fact that after subsamples in the 

 split-formalin treatment were removed from 

 whole preserved lobes, they were preserved 

 as subsamples for several weeks before the 



