676 



Fishery Bulletin 93(4), 1995 



tions (Colton and St. Onge, 1974; Smith et al., 1975; 

 Colton et al., 1979; Morse et al., 1987), show that 

 spawning occurs from April through December. There 

 is contradictory evidence for a split spawning sea- 

 son. Gonadal development (Wilk et al., 1990) indi- 

 cated that spawning off New Jersey and New York 

 peaks in May and again in September. Split spawn- 

 ing was reported to occur off Virginia and North Caro- 

 lina (Smith et al., 1975) for Long Island Sound 

 (Wheatland, 1956) and for Great South Bay, New 

 York (Dugay et al., 1989; Monteleone, 1992). How- 

 ever, other studies found no evidence for a split 

 spawning season in either Long Island Sound 

 (Perlmutter, 1939) or in ocean waters north of Vir- 

 ginia (Smith et al., 1975). In addition, Colton and St. 

 Onge (1974) collected larvae on Georges Bank from 

 July to November and found no indication of a split 

 spawning season. Spawning apparently occurs at 

 bottom water temperatures of 6-20°C (Bigelow and 

 Schroeder, 1953; Wheatland, 1956; Smith et al., 

 1975). Most spawning (70%) off Virginia and North 

 Carolina was found over bottom temperatures be- 

 tween 8.5 and 13.5°C; spawning stopped when tem- 

 peratures exceeded 15°C (Smith et al., 1975). 



This study presents analyses of the reproductive 

 seasonality of windowpane based on seasonal shifts 

 in larval abundance and on their relationships to 

 bottom water temperatures. Bottom trawl catches 

 and published accounts of distribution and abun- 



dances are used to follow the fate of the young flat- 

 fish after they settle to the bottom. Together these 

 data describe the seasonal distribution, abundance, 

 and life history of windowpane in the Middle Atlan- 

 tic Bight and on Georges Bank. 



Materials and methods 



Larvae 



Data from a variety of sources have been analyzed 

 (Table 1). Larval windowpane were collected from 

 continental shelf waters from Cape Hatteras, North 

 Carolina, to Nova Scotia during NMFS Marine Re- 

 sources Monitoring, Assessment, and Prediction 

 (MARMAP) surveys from 1977 to 1987 (Sherman, 

 1980). Surveys were conducted six to eight times each 

 year and occupied 150-180 stations (Fig. 1). At each 

 station, a 61-cm bongo net array was lowered to 

 within 5 m of the bottom or to a maximum depth of 

 200 m. Fish larvae from a 0.505-mm mesh net were 

 identified to the lowest taxon possible, enumerated, 

 measured (±0.1 mm) and rounded to the nearest 

 whole millimeter as either notochord or standard 

 length. Catches were standardized to the number of 

 larvae under 100 m 2 of sea surface on the basis of 

 the depth of the tow and the volume of water filtered 

 (Sibunka and Silverman, 1984, 1989). An additional 



