ESCAPEMENT BY FISHES FROM MIDWATER TRAWLS: 

 A CASE STUDY USING LANTERNFISHES (PISCES: MYCTOPHIDAE) 



John V. Gartner, Jr.,' Walter J. Conley,^ and Thomas L. Hopkins' 



ABSTRACT 



Escapement of fishes through the meshes of a trawl is a recognized but unquantified problem in estimating 

 size of mesopelagic fish populations. This paper provides estimates of net escapement by midwater fishes, 

 using the lanternfishes as an example. Comparison of overall catches for Tucker trawls of 1.6 mm and 

 4.0 mm mesh show highly significant differences. The small mesh size outcaught the large by a factor 

 of 2.7 for fishes smaller than 30 mm SL, while for larger fishes, the small mesh catches averaged 90% 

 of the larger mesh. Among six ranking abundant species, three patterns of escapement were observed, 

 based on significant differences in cross-sectional fish dimensions and morphological characters: 1) The 

 entire size range of the species was significantly underestimated {Benthosenm suborbitale and Notolychnits 

 valdiviae); 2) only size ranges below those of sexually mature adults were significantly underestimated 

 (Lampanyctus alatus and Lepidophanes guentheri); 3) only juveniles <30 mm SL were significantly 

 underestimated (Ceratoscopelu^ townsendi and Diaphus dwnerilii). "Conventional" midwater trawl 

 meshes of 4 to 6 mm diameter mesh provide adequate data for general distributional surveys and also 

 for some quantitative estimations such as overall biomass. Determinations of juvenile biomass, spavra- 

 ing period, trophic impact, and relative species abundances based on conventional mesh collections may 

 be prone to substantial error depending on species size. It is suggested that a net mesh of <2 mm be 

 used in conjunction with larger mesh trawls if quantitative life history data on smaller size classes and 

 species are required. 



Requisite to studies of the roles of mesopelagic 

 fishes in deep-sea ecological processes are accurate 

 determinations of species composition and the ver- 

 tical and horizontal structure of populations. Al- 

 though these are now well documented for many 

 groups in many regions of the world ocean (see Mar- 

 shall 1980), accurate abundance estimates, particu- 

 larly over the entire size range of a species, are often 

 not possible because of sampling biases related 

 to net construction and trawling methods (Stein 

 1985). 



Two factors responsible for much of the difficul- 

 ty in estimating abundance of midwater fishes are 

 net avoidance by large size classes and escapement 

 through the net meshes during capture by small 

 fishes of slender body shapes (Harrisson 1967). Both 

 result in underestimates of species abundance, 

 which can apply to either particular size classes, or, 

 in the case of diminutive species, an entire popula- 

 tion. 



While some studies show that net avoidance may 

 be reduced through the use of trawls with large 



'University of South Florida, Department of Marine Science, 140 

 Seventh Avenue S.E., St. Petersburg, FL 33701. 



^Florida Department of Natural Resources, Florida Marine Re- 

 search Institute, 100 Eighth Avenue S.E.. St. Petersburg, FL 

 33701. 



Manuscript accepted October 1988. 

 Fishery BuUetin. U.S. 87:213-222. 



mouth areas, there are a number of attendant diffi- 

 culties including enhanced escapement due to in- 

 creased mesh size (Stein 1985). Of the two problems, 

 net avoidance remains the most difficult to quan- 

 tify. Escapement is more easily calculated, but little 

 quantitative research has been directed towards this 

 problem in studies of midwater fishes (Harrisson 

 1967; Clarke 1983a). 



In this study we quantify escapement through net 

 meshes of midwater trawls using the lanternfishes 

 (family Myctophidae) as an example. The ecological 

 implications of net escapement are discussed. 



MATERIALS AND METHODS 



Myctophids were collected during eight cruises 

 aboard the RV Suncoaster from an area centered 

 at lat. 27°N, long. 86°W. The cruises covered four 

 seasons over a period of 30 months. Sampling 

 months were September (1984), November (1985), 

 January (1986, 1987), March (1985, 1987), May 

 (1986), and July (1985). Station data are presented 

 in Table 1. 



All samples were taken using modified Tucker 

 trawls fished open in an oblique "V" sweep from 

 the surface to 200 m at night. This depth range en- 

 compasses the peak nighttime abundance of all 



213 



