686 



Fishery Bulletin 101(3) 



temperature was recorded every half hour at the seamount 

 by a Stoaway Tidbit temperature logger ( Onset Computers 

 Corp., Pocassett, MA) attached to the mooring line adjacent 

 to the tag-detecting monitor. We calculated a daily tem- 

 perature by averaging the half-hourly temperatures. 



We used log-survivorship analysis (Fager and Young, 

 1978) to ascertain whether the tunas returned to the moni- 

 toring stations after favored time periods. A frequency his- 

 togram of the time intervals between randomly occurring 

 point events in a Poisson process is described by a negative 

 exponential distribution. A log-survivor plot of these inter- 

 vals generates a straight line with a slope proportional to 

 the probability of an event occurring at a given time after 

 the preceding event. This analysis is used to identify inter- 

 vals between events that occur more frequently than ex- 

 pected by chance because inflections in the resulting curve 

 are more easily distinguished from a .straight line than the 

 shape of the distribution on a frequency histogram with a 

 negative exponential distribution. An inflection in the log- 

 survivor curve indicates a change in the probability of an 

 event occurring at a given time after the last event — in our 

 case the time between successive arrivals of tunas within 

 the ranges of the two monitors. 



Results 



Twenty-three yellowfin tunas were tagged from 11 April 

 1998 to 12 September 1998 (Table 1). Individuals were 

 tagged during daylight hours from 6:41 to 19:07 hours. 

 The tunas ranged in length from 71.0 to 155.0 cm TL. 

 They ranged in mass from 7.25 to 73.1 kg. There appeared 

 to be two discrete size classes, small individuals varying 

 from 7.25 to 23.0 kg and large ones from 51.7 to 73.1 kg. 

 The masses of the larger individuals were determined 

 from their lengths by using a regression equation (Moore, 

 1951). 



The yellowfin tunas stayed at seamount Espiritu Santo 

 over varying time periods (Fig. 2). Nine of the 23 tunas 

 left the seamount on the same day that they were tagged 

 (Fig. 2A). Two of the nine returned to the seamount once 

 for a single day. one within a week of tagging and the 

 other after two and one-half months. Six tunas stayed 

 intermediate periods of time after tagging, ranging from 

 two to six weeks. One of these tunas (no. 9) was eventually 

 caught at the seamount. Another tuna (no.lO) visited for 

 a single day after an absence of five weeks and returned 

 again after a similar period to stay for 15 months. Four 



