462 



Fishery Bulletin 91 [3). 1993 



seamounts in the following year. Allowing a one-year 

 period (recruitment by the following summer) for indi- 

 viduals of the 1+ epipelagic year-class to move from 

 the open ocean site (45°N 155°W) to the SE-NHR sea- 

 mounts and also that the total distance traversed is 

 twice the straight-line distance between these sites 

 (assuming this is comparable to results of open-sea 

 tracking of coho salmon Oncorhynchus kisutch reported 

 in Ogura and Ishida (1992)), a rough estimate of the 

 average ground speed required is 18 cm/s if currents 

 are disregarded. Based on a length (27.1cm FL) mid- 

 way between the average FL of 1+ (x = 24.49 cm FL) 

 and 2+ year-old (x = 29.64cm FL) epipelagic speci- 

 mens collected during summers at the open ocean site, 

 individuals would need to swim 0.66 body length/s to 

 arrive at an SE-NHR seamount one year later. This 

 speed is close to the mean sustained cruising speeds 

 (0.2-0.5 body length/second) typical of horizontally mi- 

 grating fish (Beamish, 1978). Since no evidence exists 

 of gonadal maturation in the epipelagic stage 

 (Humphreys et al., 1989), the major energy demand 

 during seamount migration is likely locomotion, in ad- 

 dition to basal metabolism. It is unknown whether 

 feeding occurs during this migration; however, 

 McKeown (1984) suggests that even if fish species feed 

 during migration, the added energy intake may be ne- 

 gated by the additional energy required for feeding 

 activity. Regardless of feeding, lipid reserves are typi- 

 cally used during migration (McKeown, 1984), and their 

 storage in the liver and muscle makes these prime 

 sites from which energy reserves can be drawn 

 (Woodhead, 1975). As such, the liver is susceptible to 

 net decreases in weight by the end of migration; an 

 example of this occurs among migrating Barents Sea 

 cod Gadus morhua (Woodhead, 1975). Evidence of fewer 

 prey contents among higher FI armorhead examined 

 at SE Hancock 3 indicates that recent recruits initially 

 may feed little at the seamounts. This suggests that 

 the liver remains susceptible to continued energy deple- 

 tion and therefore continued decrease in liver weight. 

 Assuming that liver weight is declining relatively faster 

 than body weight, the HSI will not only decrease by 

 the time of initial recruitment to the seamount but 

 perhaps also during the interval between initial re- 

 cruitment and sampling. Therefore, an annual or sea- 

 sonal change in mean HSI among new recruits may be 

 caused by inter- or intra-annual differences experienced 

 by epipelagics in terms of environmental conditions 

 and distance traversed during the seamount migra- 

 tion, along with an unknown time gap between initial 

 recruitment and sampling. Hence, using HSI as a re- 

 cruitment indicator appears to be inherently more 



'M. P. Seki, NMFS Honolulu Laboratory, pers. commun. January 

 1992. 



labile and thereby less feasible and reliable than 

 the method based on the absence of mature M. 

 macropharynx. 



The parasite approach revealed that new recruit- 

 ment coincided primarily with the late spring and mid- 

 summer sampling periods. Somerton and Kikkawa 

 (1992) have analyzed the recruitment patterns of 

 armorhead to SE Hancock, using the modal progres- 

 sion of FI in which separate modes are considered an- 

 nual cohorts. If a modal influx of high FI fish repre- 

 sents a "recruitment cohort," estimated recruitment 

 was highest in terms of biomass (58 1) and percentage 

 (79%) of the SE Hancock population during August 

 and October 1986, whereas the next highest recruit- 

 ment (8t and 21% of the population, respectively) in 

 July and August 1988 was comparatively modest. In- 

 terestingly, however, the level of new recruitment, as 

 determined by the percentage of sampled armorhead 

 (FI >0.26) un-infected with mature M. macropharynx, 

 was highest in June 1985 and very low in October 

 1986. These results appear contrary to stock assess- 

 ment based estimates of modest and high proportions 

 of recruit-of-the-year armorhead (FI >0.26) in the sea- 

 mount population during June 1985 and October 1986, 

 respectively (see Fig. 2). These results can be recon- 

 ciled if we consider that the rate of post-recruitment 

 decline in FI (estimated at 0.00169/month; Somerton 

 and Kikkawa (1992)) does not allow one to decipher 

 between recruits which may have just arrived and those 

 which arrived months prior to sampling. The absence 

 of mature stage M. macropharynx appears to be the 

 only criteria examined in this study capable of detect- 

 ing whether a recruit-of-the-year (FI >0.26) is a new 

 arrival to the seamount. Based on this criteria, the 

 level of new recruitment during the months sampled 

 was highest in June, less so in July and August, and 

 very low during October, November, January, and April. 

 These results tend to corroborate evidence of similar 

 seasonal armorhead recruitment at other SE-NHR sea- 

 mounts (Boehlert and Sasaki, 1988). 



Recruits identified via the parasite approach typi- 

 cally have FI values >0.26 (x = 0.291). However, a high 

 FI value alone provides no assurance that an indi- 

 vidual is a new recruit. The difference in mean FI 

 between epipelagic individuals and new recruits (un- 

 infected with mature M. macropharynx) may indicate 

 that FI decreases sometime before epipelagics of re- 

 cruitment size reach the seamounts. This can be ad- 

 equately examined only after both groups are further 

 sampled. 



Acknowledgments 



We are indebted to the officers and crew of the NOAA 

 ship Townsend Cromwell, the RV Oshoro Maru, and 



