170 



Fishery Bulletin 102(1) 



To characterize prowfish distribution we obtained catch 

 data from 42,601 bottom trawl deployments (hauls) exe- 

 cuted from 1953 through 2000 using a variety of net de- 

 signs. We used these data to determine presence or absence 

 of prowfish at each haul location. Previous observations 

 have indicated that prowfish tend to be pelagic as larvae 

 and become demersal as adults (Matarese et al., 1989; 

 Hart, 1973). A full accounting of prowfish distribution by 

 life stage is beyond the scope of this investigation, which 

 focuses on adults. Therefore we confined our observations 

 to haul catches taken on bottom, as opposed to in mid-water 

 or at the surface. 



On two of the bottom trawl surveys, one in the Gulf of 

 Alaska from 22 May to 30 July 1996, and the other in the 

 Aleutian Islands from 10 June to 11 August 1997, addi- 

 tional prowfish data were collected. Consistency between 

 these surveys in sampling procedures and equipment 

 (Martin, 1997, and Stark 3 ) facilitated subsequent data 

 comparisons. 



Density of prowfish at each sampling location was 

 estimated as the number caught divided by the km 2 of 

 area swept by the trawl (catch per unit of effort, or CPUE). 

 Research vessels on both surveys employed the standard 

 RACE Division model poly-Nor'eastern high-opening bottom 

 trawl net with roller gear, and hauls were made during day- 

 light. Net configuration and bottom contact during trawling 

 were monitored by Scanmar instrumentation. Data were ob- 

 tained from 807 hauls in the Gulf of Alaska and 408 hauls in 

 the Aleutian Islands. The average area swept per haul was 

 0.025 km 2 in the GOA and 0.024 km 2 in the AI. 



All prowfish were sorted to sex by examination of the 

 gonads and then length (total length; cm) was measured. 

 Sample sizes were 84 males and 90 females for the Gulf 

 of Alaska; 396 males and 431 females for the Aleutian 

 Islands. Whole-body weights (g) of 83 male and 88 female 

 prowfish from the Gulf of Alaska were measured and the 

 sagittal otoliths were removed and stored in 50% ethanol. 

 Whole ovaries from a representative subsample of 39 of 

 the females were removed, frozen, and later stored in 10% 

 buffered formalin solution. 



Diet composition was examined from stomach contents of 

 76 individuals (18 from the Gulf of Alaska and 58 from the 

 Aleutian Islands). Stomachs containing food and with no 

 signs of regurgitation or net-feeding (e.g. the stomach was 

 in an inverted or flaccid state or there was the presence of 

 prey in the mouth or around the gills) were removed and 

 preserved in 10% buffered formalin. 



Laboratory procedures 



Standard otolith-prcparation techniques for age determi- 

 nation were modified to accommodate the relatively small 

 size of prowfish otoliths (usually <5 mm long). An anterior 

 portion of each otolith was removed by a transverse cut with 

 scalpel perpendicular to the sagittal axis and anterior to the 



:! Stark, J. 1998. Report to industry: fishing log for the 1997 

 bottom trawl survey of the Aleutian Islands. AFSC Proc. Rep. 

 98-06, 96 p. Alaska Fish. Sci. Cent., Natl. Mar. Fish. Serv., NOAA. 

 7600 Sand Point Way NE, Bldg. 4, Seattle, WA 98115-0070. 



nucleus. The remainder, which contained the nucleus, was 

 baked at 300-475°C for up to 17 min or heated over an alco- 

 hol flame to enhance visibility of annuli. The otoliths were 

 then individually mounted on slides by completely embed- 

 ding them in clear thermoplastic posterior end down. On 

 hardening, each mount was wet-sanded on increasingly fine 

 grades of sandpaper (400-2000 grit), parallel to the slide, 

 until the surface intersected the otolith nucleus (trans- 

 verse section). Preparing readable mounts was a delicate 

 procedure; besides cutting and polishing the small otoliths 

 precisely without fracturing them, precise heating tem- 

 perature and time were especially critical to expose annuli 

 without again causing fractures or burning the otolith. 

 Our method had advantages over the standard "break and 

 burn" method of simply coating the surface of a temporarily 

 mounted specimen with oil to enhance visibility of annuli, 

 in lieu of polishing. It allowed a more precise intersection of 

 the nucleus by the viewed surface and eliminated the need 

 to remove oil from specimens intended for further viewing 

 in order to prevent blurring of annuli. After preparation, 

 slides were placed in sufficient water to cover the surface 

 scratches and were examined under a dissecting microscope 

 with reflected light. Age in years was determined by count- 

 ing the annuli or hyaline bands according to the criteria 

 described in Chilton and Beamish (1982). 



Prowfish ovaries were prepared for histological examina- 

 tion by removing a small portion from the middle of each 

 ovary, which was then embedded in paraffin, sectioned at 

 6 jim, and stained with hematoxylin and eosin. The histo- 

 logical slides were examined under a compound microscope 

 and donor females were classified as either sexually imma- 

 ture or mature based on the presence of yolk in the oocytes 

 (i.e. vitellogenesis). 



Prowfish stomachs were processed by first neutralizing 

 the 10% formalin used for initial fixation and then by im- 

 mediately transferring the stomachs into 70% ethanol. The 

 food was removed, blotted with a paper towel, and exam- 

 ined with a dissecting microscope. Prey items were sorted 

 to the lowest practical taxonomic level and then weighed 

 to the nearest 0.1 gm. The percentage of total prey weight 

 which each taxon comprised, as well as the percentage of 

 stomachs containing each taxon, was calculated for each 

 haul sample and then estimated for each of the two regions 

 as the average of the per-haul percentages. 



Analysis of data 



The distribution of prowfish density over depth in the 

 Gulf of Alaska and the Aleutian Islands was determined 

 by calculating the mean CPUE for each 20-m depth inter- 

 val from 20 m to 480 m. Both surveys utilized a stratified 

 sampling design in which sampling density (hauls per 

 unit area) varied by geographical subarea (Martin, 1997; 

 Stark 1 ). To compensate for this variation, the CPUE of each 

 haul was weighted by the inverse of the sampling density 

 in that geographic stratum. The mean bottom depth as 

 weighted by prowfish density was calculated for each of 

 the two regions as the weighted average of the midpoints 

 of the depth intervals, where the weighting factors were 

 the interval-mean CPUE values. 



