Smith et al.: Distribution and biology of Zaprora stlenus 



177 



lands area. Because of the lack of a relatively broad shelf 

 in the region, a larger proportion of trawls are in or near 

 areas of steep seafloor gradient and therefore likely over 

 rough bottom (Fig. 2). 



Length distribution 



In both the Gulf of Alaska and the Aleutian Islands, few 

 prowfish <40 cm in length were captured (Fig. 4). This 

 paucity of small prowfish is not due to size selection by 

 the trawl net mesh because the codend is lined with small 

 mesh (1.3 cm stretched measure) webbing that retains 

 small individuals of other species. A different explanation, 

 based on the observations of Brodeur (1998) and Scheffer 

 (1940), is that pre-adult prowfish are pelagic, remaining 

 in proximity with large coelenterates and thus avoiding 

 bottom trawls. Thus, the minimum capture length may 

 indicate the length at which prowfish recruit to a demer- 

 sal habitat. Our data showed no statistically significant 

 length difference between sexes, in contrast with the data 

 of Tokranov (1999) who suggested a length dimorphism 

 where females are generally longer than males. 



Weight-length relationship 



The best-fitting model of weight versus length predicts 

 that for any length, female prowfish are, on average, 3.7% 

 heavier than males (Fig. 5). It seems unlikely that this 

 relationship exists over all developmental stages because 

 our samples were almost all adults and such a (relative) 

 difference might not remain constant during all ontoge- 

 netic sexual divergence. What is more certain is simply 

 the existence of some small degree of length-weight 

 dimorphism (females slightly heavier at a given length). 

 Also, this dimorphism is not likely to stem primarily from 

 a sexual difference in gonad weight because the maximum 

 proportion of total female body weight composed of ovarian 

 tissue was only 2.7%. Thus the difference is due to other 

 morphological or behavioral differences. 



Growth 



There was no significant difference between sexes in length 

 versus age. The predicted length of a prowfish of given age 

 based on our samples was higher than that indicated by 

 Tokranov (1999). In our study 5-year-old and 9-year-old 

 fish averaged 56.6 cm and 73.5 cm in length, respectively. 

 Tokranov (1999) considered that prowfish growth deter- 

 mined from otoliths of 102 specimens from the Northwest 

 Pacific indicated a comparatively fast-growing species 

 reaching an average length of 44.6 cm by 5 years of age 

 and 67 cm after 9 years. These data suggest prowfish are 

 indeed relatively fast growing and that growth rates for the 

 Gulf of Alaska are faster than those for off southeastern 

 Kamchatka and the northern Kuril Islands. Alternatively, 

 size-dependent mortality from such elements as incidental 

 capture by commercial fishing may affect the two popula- 

 tions differently. 



Historically, two other prowfish have been aged from 

 otoliths: a male 84 cm long taken near Eureka, CA (Fitch 



and Lavenberg, 1971), and a female 50.1 cm long ( standard 

 length) from off Monterey (Cailliet and Anderson, 1975). 

 The ages estimated were 12 and 3 years, respectively. Af- 

 ter converting the standard length record to an estimate 

 of total length for the second specimen of 58 cm by using 

 a ratio described by Baxter, 6 both lengths are slightly 

 greater than our predictions for the same ages, albeit near 

 the limits of our data range. This finding contrasts with 

 the predictions of lesser length at a given age presented 

 by Tokranov (1999). 



Maturity 



Little previous data exist with which to compare our obser- 

 vations of female prowfish rate of maturation. Cailliet and 

 Anderson (1975) examined the ovaries of their 50.1-cm 3- 

 year-old female specimen for vitellogenesis and predicted 

 an age at first spawning of 4 years, slightly less than the 

 lower 95% confidence limit of 4.4 years for our expected 

 average age at 50% maturity. 



Food habits 



Our observation that gelatinous zooplankton was the 

 largest constituent in the contents of prowfish stomachs 

 (Table 1) is supported by Tokranov ( 1999), who found that 

 the two most common prey taxa among the contents of 102 

 stomachs of adult specimens from the northwestern Pacific 

 were Scyphozoa (59.6-62.0% of stomachs) and Ctenophora 

 (6.0-15.4% of stomachs). Anecdotal observations have also 

 been made of the feeding behavior of an aquarium specimen 

 over an approximate 2-year period (Carollo and Rankin, 

 1998). When first obtained, the fish ate only various jel- 

 lyfish species, rejecting other food, including a variety of 

 live invertebrates. In our food samples, we observed other 

 taxa, such as invertebrates and small fish, but these were a 

 minor part, possibly first captured by jellyfish and then sec- 

 ondarily ingested by prowfish. Carollo and Rankin (1998) 

 found that the aquarium specimen would ingest such items 

 when eating the bells of Chrysaora melanaster in which 

 such food had previously been placed, indicating the pos- 

 sibility of this occurring naturally. Possibly more reflective 

 of the unnatural circumstances, the specimen later began 

 accepting such items outside the bells of jellyfish. 



Apparent adaptations of the prowfish to a diet of ge- 

 latinous zooplankton include the small, sharp, close-set 

 teeth in a single row attached only to the jaws, which are 

 capable of a 180-degree gape, and the large rough-scaled 

 lips (Clemens and Wilby, 1961; Hart, 1973; Carollo and 

 Rankin, 1998). 



Acknowledgments 



We are grateful for the expert advice given by Alaska Fish- 

 eries Science Center colleagues Delsa Anderl and Nancy 



6 Baxter, R. 1990. Unpubl. manuscript. Annotated key to the 

 fishes of Alaska, 803 p. [Available from Sera Baxter, Box 182, 

 Seldovia, AK 99663.1 



