540 



Fishery Bulletin 105(4) 



oocytes removed from the same localized population 

 of oocytes all had the same mean size and density. 

 Each of these subsamples was then preserved in one 

 of four ways, such that one subsample from each fish 

 was placed in each treatment (i.e., one replicate per 

 fish within a treatment). Only subsamples from speci- 

 mens collected in May in the GOM (i.e., 23 cod and all 

 16 plaice) were weighed to the nearest 0.001 g before 

 preservation. 



In the formalin treatment (termed the sub-formalin 

 treatment in weight comparisons to distinguish it from 

 the lobe-formalin treatment), a subsample of ovarian 

 material was placed in a vial (10-mL plastic vial with 

 screw cap) containing 5 mL of formalin. Similarly in the 

 Gilson's treatment, a subsample of ovarian material was 

 placed in a vial containing 5 mL of modified Gilson's 

 solution. This modified form of Gilson's solution was 

 employed because it does not contain mercuric chloride, 

 and is thus not toxic, but still effectively preserves and 

 separates oocytes (Friedland et al., 2005). These sam- 

 ples were capped and shaken to ensure oocytes were all 

 thoroughly immersed in the preservative, and were then 

 stored upright in a rack at room temperature. In the 

 ethanol treatment, a subsample of ovarian material was 

 placed in a vial and immersed in 4 mL of 95% ethanol 

 for approximately 15 seconds, before 1.5 mL of distilled 

 water was added to the sample through a graduated 

 pipette, before the solution was diluted to 70% ethanol. 

 Samples in this treatment were also shaken and stored 

 at room temperature. Briefly submerging the material in 

 95% ethanol is meant to act as a fixative, whereas 70% 

 ethanol is better for long-term preservation (Black and 

 Dodson, 2003). In the freezing treatment a subsample 

 of ovarian material was immersed in a vial of distilled 

 water. Because of our concern that the expansion with 

 freezing too much water might crack the vials, 3 mL of 

 water was used rather than 5 mL as in the other treat- 

 ments. These samples were also shaken, but then placed 

 (upright) in the freezer so that the ovarian material 

 would be frozen solid in ice. The samples were frozen in 

 this way to preserve the shape of the oocytes and to stop 

 them from drying out or becoming freezer damaged. 



Subsamples that were weighed fresh (i.e., those col- 

 lected in May) were reweighed after 97-111 (mean=102) 

 days of preservation and then returned to their pre- 

 servative vials. To weigh a preserved subsample, the 

 entire content of a vial was poured into a 40-^( mesh, 

 nylon cell-strainer fitted atop a hand operated vacuum 

 pump. Excess fluid was then removed from each sample 

 by repeatedly squeezing the handle of the hand pump 

 until preserved ovarian material was all that remained 

 in the strainer. The strainer and its contents were then 

 weighed to the nearest 0.001 g, and the known weight 

 of the strainer was subtracted to find the preserved 

 subsample weight. 



In addition to the four treatments mentioned above, a 

 fifth split-formalin treatment (so termed because pres- 

 ervation was split into two phases described below) was 

 used for the left ovarian lobes. After 91-131 (mean=lll) 

 days of preservation in formalin, each left lobe was 



removed from its jar and a 1.5-mL subsample was re- 

 moved from its center and placed in a vial containing 

 5 mL of formalin. Because they were not preweighed, 

 these subsamples were used only for examining pre- 

 servative effects on oocyte size. These subsamples re- 

 mained in vials for another 33-92 (mean = 63) days 

 before they were analyzed. Including the total time 

 that split-formalin samples were preserved, the time 

 that samples from all treatments were preserved was 

 117-164 (mean=145) days before the mean oocyte diam- 

 eter of each sample was estimated. 



The method used here for measuring oocytes is largely 

 based on the work of Thorsen and Kjesbu (2001) but is 

 described here because of differences in details. To start, 

 a vial was shaken vigorously for 30-60 seconds to break 

 apart any oocytes still connected. Samples in modified 

 Gilson's solution did not require shaking because the 

 oocytes were already chemically separated. Most other 

 samples broke apart very well from this shaking, es- 

 pecially those with larger oocytes. Samples in ethanol 

 and samples containing the smallest oocytes were more 

 difficult to break apart and required more shaking. 

 After the vial was shaken, a transfer pipette was used 

 to agitate the oocytes by the action of drawing in and 

 expelling the solution rapidly, so that all oocytes were 

 effectively mixed randomly in suspension and would 

 not settle out by size. While the suspension was being 

 agitated, a portion of the vial's contents was drawn and 

 quickly emptied into a Petri dish containing =2 mm of 

 the respective preservative and a drop of 5% Palmolive 

 soap solution which helped the oocytes spread out in the 

 dish and kept them from floating. Oocytes were then 

 added to the Petri dish until the bottom of the dish was 

 filled but so that the oocytes could still be spread out 

 without clumping. A black lid was then placed on the 

 dish to serve as a contrasting background, and an image 

 of the sample was captured with a flatbed scanner. Each 

 sample was scanned at 1200 dpi in 16-bit gray scale and 

 a contrast setting of 18 with the use of an Epson Perfec- 

 tion 1670 scanner (Epson, Long Beach, CA). The same 

 selection marquis was used for each sample so that all 

 images were exactly the same size: 3688x3671 pixels, 

 within a 12.9 MB uncompressed TIFF file. 



Because samples in most treatments were stored at 

 room temperature, no temperature adjustment was need- 

 ed before they were scanned. For samples in the freez- 

 ing treatment, however, care was taken to be sure that 

 vials were thawed one hour before they were scanned 

 out of concern that oocyte diameter may be affected 

 by how much time thawed samples were left in water. 



Once scanned, each image was analyzed in the free- 

 ware program Scion Image (Scion Image, version beta 

 4.0.2, Scion Corporation, Frederick, MD) by first setting 

 the scale of each image to 1200 pixels per inch (472 

 pixels per cm), and the measurements to micrometers. 

 The lower limit of the density slice tool was then set to 

 40 and the upper limit ranged from 180 to 205, and was 

 typically 195. The upper limit varied because the grey 

 value of the oocyte margins varied between samples 

 depending on how dark the oocytes were. The Scion 



