Rotunno and Cowen: Temporal and spatial spawning patterns of Peprilus triacanthus 
795 
the number of caudal vertebrae and ventral midline 
melanophores. 
Age and growth 
Ages of larval and juvenile P. triacanthus can be de- 
termined by counting otolith increments. Validation 
was necessary because of the prevalence of subdaily 
increments in this species. Secondary nuclei 
(multinucleation) were also noted in Peprilus otoliths. 
The cause and timing of the formation of secondary 
nuclei are not presently understood (Campana and 
Neilson, 1985), although these secondary nuclei have 
been demonstrated to form during metamorphosis 
in bluefish (Hare and Cowen, 1994). Secondary nu- 
clei, in butterfish otoliths, do not seem to follow a 
consistent pattern in the development of the fish; we 
found that two sagittal bones from one fish frequently 
contained differing numbers of secondary nuclei. Al- 
though secondary nuclei may form during metamor- 
phosis (Campana and Neilson, 1985; Hare and 
Cowen, 1994), we found secondary nuclei in larvae 
that were several millimeters smaller than the size 
at which metamorphosis occurs (16 mm). 
Larval and early juvenile P. triacanthus from 6.0 
to 28.0 mm SL grew at a rate of 0.227 mm/day. Al- 
though ages of young butterfish have not been re- 
corded previously, Colton and Honey (1963) gave sizes 
of P. triacanthus from hatching to six days of age. 
Based on their estimates, growth rates ranged from 
0.01 to 0.55 mm/day and decreased with the age of 
the fish. Specimens of young P. burti have been aged 
with modal length-frequency analysis; growth rates 
of P. burti ranged from 0.25 to 0.56 mm/day (Murphy 
and Chittenden, 1990). Peprilus burti may be ex- 
pected to have a higher growth rate than P. 
triacanthus owing to the fact that it spawns in 
warmer waters of the Gulf of Mexico. 
Our analysis of hatching-date distributions dem- 
onstrated that P. triacanthus has a more extensive 
spawning season than previously reported. Their 
spawning effort seems to be focused into two cohorts 
(spring: February-March; and summer: June-July) 
although evidence is presented that suggests at least 
some fish spawn during the interim period between 
seasonal peaks (i.e. late April to early June). Kawa- 
hara 2 speculated that P. triacanthus may begin 
spawning in April. It should be noted that Kawahara 
( 1977) based his spawning estimate on adult growth 
2 Kawahara, S. 1977. Age and growth of butterfish, Peprilus 
triacanthus (Peck), in ICNAF Subarea 5 and Statistical Area 6. 
ICNAF Res. Doc. 77/VI/27. June 1977 Annual Meeting. Far Seas 
Fishery Laboratory, Shimizu, Japan. 
rates, which are higher than our estimate for larvae 
and juveniles and, therefore, would have underesti- 
mated the spawning duration. 
A bimodal hatching-date distribution in P. tria- 
canthus is similar to that reported for another north- 
south migrating species within the western Atlan- 
tic, the bluefish, Pomatomus saltatrix (Kendall and 
Walford, 1979; Nyman and Conover, 1988). However, 
Hare and Cowen (1993) and Smith et al. (1994) have 
proposed that P. saltatrix spawns continually dur- 
ing its north-south migration and that the apparent 
bimodal hatching date distribution may result from 
advective processes acting on the larval distributions 
and from sampling artifact. 
The presence of an apparent bimodal spawning in 
P. triacanthus , with spring and summer peaks, may 
be an artifact of our sampling. Because we did not 
sample from April to June in each year and because 
our sampling locations were spatially distinct be- 
tween spring and summer, it is possible that we did 
not collect larvae spawned during May and early 
June. However, we could have collected older fish in 
our samples that were spawned early in June. Data 
collected monthly by the National Marine Fisheries 
Services (NMFS) as part of their Monitoring, Assess- 
ment, and Prediction (MARMAP) surveys suggest 
that larvae are indeed present from April to August 
in the MAB, thus it is likely that spawning occurs 
continually (Fig. 8). Moreover, MARMAP data indi- 
cate a northward progression of larvae from near 
Cape Hatteras during March or April (or both) into 
the entire MAB by mid-summer and until October. 
This spatio-temporal pattern is consistent with the 
possibility of spawning associated with a seasonal 
northward migration of adult P triacanthus. Horn 
( 1970) has speculated that butterfish movements are 
highly influenced by temperature (and salinity to a 
lesser degree). Temperature-related movements by 
butterfish (and bluefish) would correspond with the 
observed northward progression of young larvae from 
the SAB in the spring into the MAB in the summer 
in association with seasonal warming. The extent of 
north-south migration by P. triacanthus , however, 
requires further study. 
In conclusion, this study adds to our current knowl- 
edge of the early life history and spawning seasonal- 
ity of butterfish, P. triacanthus. Our finding of a more 
protracted spawning season and of a seasonal differ- 
ence between spawning locations should be of value 
in reassessing management plans for this species. 
Current management plans are based on conclusions 
that P. triacanthus spawns during summer months 
only. The apparent similarity of spawning periodic- 
ity of butterfish to that of bluefish (Cowen et al., 1993; 
Hare and Cowen, 1994, Smith et al., 1994) suggests 
