458 



Fishery Bulletin 105(4) 



database for groundfish surveys in the Gulf of Maine 

 and the Middle Atlantic Bight (Steimle et al. 1999). 

 Steimle et al. (1999) grouped all fish <43 cm TL into 

 the juvenile category. As a result, several age classes 

 were combined and there was no resolution for habitat 

 of YOY goosefish. Prior descriptions of other aspects of 

 the life history and ecology of YOY goosefish include 

 associations of fish with depth and substrate type for 

 Canadian waters (Jean, 1965; Scott, 1982), and with 

 food habits for juveniles and adults (Sedberry, 1983; 

 Armstrong et al., 1996). 



The gaps in our knowledge about YOY goosefish, as 

 articulated by Steimle et al. (1999), motivated the ef- 

 forts reported here. Our objectives were 1) to describe 

 more fully the morphological development of goosefish 

 during the transition from pelagic larvae to benthic 

 juveniles, 2) to estimate the timing, sizes, and ages of 

 early life history events, and the growth rates of YOY 

 goosefish, 3) to determine the distribution of pelagic and 

 benthic juveniles in time and space, and 4) to identify 

 the food habits and habitats of settled YOY goosefish. 



Materials and methods 



Two species of Lophius are found in the waters of the 

 northwestern Atlantic, goosefish (L. ame/'icarius Valenci- 

 ennes) and blackfin goosefish (L. gastrophysus Miranda- 

 Ribeiro) (Caruso, 2002), and thus correct taxonomic 

 identification is essential. Previously, L. gastrophysus 

 was considered to be distributed from Brazil to Flor- 

 ida, and L. aniericanus was considered to range from 

 Nova Scotia to Cape Hatteras (Bigelow and Schroeder, 

 1953; Scott and Scott, 1988) and to overlap occasionally 

 between Florida and Cape Hatteras (Caruso, 1983). Pre- 

 liminary data from the NMFS Commercial Cooperative 

 Research goosefish survey indicate that L. gastrophysus 

 has a much broader range, which is believed to extend 

 into the Gulf of Maine (Richards'). To ensure proper 

 identification of our specimens, we used the diagnos- 

 tic characters reported for these congeners (Caruso, 

 1983). 



Morphological development 



For some species of fish, ontogenetic state of an indi- 

 vidual has proven to be a better metric of early life his- 

 tory events than has age (Policansky, 1983; Fuiman et 

 al., 1998). For this reason, we used morphological and 

 meristic data from pelagic and benthic individuals of L. 

 americanus as a basis for describing early life history 

 stages and for estimating size and age at settlement. 

 We define planktonic juveniles as those individuals that 

 have completed fin ray formation (>12 mm TL; Everly, 

 2002), but have not yet settled. Benthic juveniles include 



' Richards, A. 2004. Personal comniun. National Oceanic 

 and Atmospheric Administration, National Marine Fisher- 

 ies Service, Northeast Fisheries Science Center, 166 Water 

 Street, Woods Hole, MA 02543-1026. 



all postsettlement YOY and juveniles of older age classes. 

 A total of 88 fish (9.8-188 mm TL) were collected and 

 examined (three fish per 5-mm size class) from a vari- 

 ety of sources (Table 1). These specimens were either 

 collected and frozen, preserved in a solution of 70-95% 

 ethanol or 5% formalin onboard the sampling vessel, 

 or were obtained from museum collections. Shrinkage 

 was determined by measuring fish upon capture and 

 after preservation in ethanol or formalin. In total, 28 

 morphometric and meristic characters were examined 

 (Table 2). These characters were based largely on those 

 general character definitions provided by Hubbs and 

 Lagler (1958). Modifications to these definitions in our 

 study were due to the unique features possessed by 

 L. americanus. Head length was measured from the 

 anterior tip of the premaxillary bone to the gill open- 

 ing. The length of the first dorsal spine, or illicium, 

 was measured from its base to the base of the esca, the 

 fleshy distal pendant of the illicium. The length of the 

 esca was measured from its base to its distal end. An 

 ocular micrometer was used for lengths <10 mm and a 

 dial caliper was used for lengths >10 mm; all measure- 

 ments were taken to 0.1-mm resolution. 



Changes in body proportions that occurred during the 

 pelagic-benthic transition were determined by examin- 

 ing the relation between each morphological character 

 and total length. Such changes in allometry were de- 

 tected by alteration in the slope of the line relating a 

 focal character and total length (or any other reference 

 character) on a log-log scale. When a fish is growing but 

 not changing shape (i.e., exhibits isometric growth), the 

 slope of this relationship is unity. When an individual 

 or, as used here, an ontogenetic series of individuals, 

 exhibits changes in shape with increasing total length, 

 this means that the character(s) defining shape have 

 changed in relation to total length. This transition in 

 the degree of allometry will be reflected by a systematic 

 deviation of the slope in the bivariate character plot. 

 The magnitude of the slope before and after the transi- 

 tion, as well as the total length at which the transition 

 occurred, were estimated by using piecewise regression 

 (Toms and Lesperance, 2003). 



Unlike measures of body proportions, other characters 

 that change during the larval- juvenile-adult states 

 are difficult to quantify. For example, the degree of 

 pigmentation and the development of tubercles (Wiley 

 and Collette, 1970) and cirri (fleshy flaps or tags, see 

 Caruso, 2002) change gradually and in multiple dimen- 

 sions during the progression from larval to juvenile 

 phenotypes. These types of characters were scored as 

 or 1 for the larval state and juvenile-adult state, re- 

 spectively. The juvenile-adult state was assumed to be 

 represented by the largest size class that we examined 

 (185-190 mm TL). For each of these characters, the 

 median and standard error (SE) of size at transition 

 to the juvenile/adult state were identified by fitting a 

 cumulative normal distribution to the 0-1 scores by us- 

 ing a probit regression model. Because these estimates 

 are based on data at the population level, our estimate 

 of the mean size at transition was actually the size 



