816 
Fishery Bulletin 95(4), 1 997 
spawning stocks have been supplemented by hatch- 
ery releases, but scale-pattern analysis and coded- 
wire tag recoveries of spawning fish intercepted at 
the fence allow each run to be partitioned into hatch- 
ery-reared and naturally spawning components 
(Unwin and Glova, 1997). The stability of the flow 
regime ensures that pre-emergence mortality of ova 
and alevins is not flow-dependent and is reflected by 
the lack of interannual variation in egg to fry sur- 
vival for naturally spawning fish (Unwin, 1986; 
Unwin, in press). Over five years (1973-76, and 1992) 
of record egg-to-fry survival ranged from 38 to 52% 
and averaged 48%. On this basis, annual production 
can be consistently expressed in terms of the num- 
ber of fry leaving Glenariffe Stream (Unwin, 1997). 
For this study, I used the data set in Table 1 of 
Unwin (1997), summarizing fry-to-adult survival (S) 
for the 26 years from 1965 to 1990, expressed as live 
adult spawners reaching Glenariffe Stream (summed 
over all year classes for each cohort) per 10,000 fry. 
These data are reproduced here as Table 1. Survival 
ranged from 1.3 (for the 1971 brood year) to 117 (for 
the 1973 brood year), with an annual mean of 8 
spawners per 10,000 fry (0.079%). These data were 
log-normally distributed (Unwin, 1997); therefore I 
used log-transformed values for all calculations (see 
also Bradford, 1995). 
Data analysis 
For each year from 1965 to 1990, 1 calculated mean 
flows ( Q ) for each calendar month, for the two three- 
month periods August-October (“spring”) and No- 
vember- January (“summer”), and for the full six 
months. As indices of flow variability, I determined 
the maximum flow ( Q ), and the ratio of the mean to 
the median flow ( Q ), for the same monthly, three- 
monthly and six-monthly intervals. Although other 
measures of flow variability are possible, such as 
skewness, coefficient of variation (CV), and baseflow 
index (which measures the ratio of the volume of base 
flow to the volume of total runoff), these all tend to 
be highly correlated and there is no one measure 
which represents the “best” index (Jowett and 
Duncan, 1990). I chose Q because it is more robust 
(i.e. insensitive to extreme outliers) than statistics 
Table 1 
Spawning population size, fry production, adult returns, and fry-to-adult survival (adults per 10,000 fry, rounded to the nearest 
integer) for naturally spawning Glenariffe Stream chinook, 1965-90. Fry production for 1973-76 was estimated from trapping 
records; all other figures are based on a mean egg-to-fry survival of 48% (Unwin, 1997). 
Brood 
year 
Number of 
female spawners 
Estimated fry 
production (thousands) 
Number of 
returning adults 
Adults per 
10,000 fry 
1965 
1,278 
2,988 
4,676 
16 
1966 
573 
1,513 
1,334 
9 
1967 
746 
1,760 
505 
3 
1968 
1,781 
4,310 
2,642 
6 
1969 
1,286 
3,249 
2,760 
8 
1970 
248 
655 
474 
7 
1971 
1,084 
2,573 
330 
1 
1972 
1,618 
3,418 
1,731 
5 
1973 
160 
275 
3,207 
117 
1974 
173 
426 
1,242 
29 
1975 
799 
1,834 
2,045 
11 
1976 
1,522 
3,436 
2,943 
9 
1977 
778 
1,614 
1,341 
8 
1978 
863 
1,810 
1,787 
10 
1979 
1,413 
2,138 
546 
3 
1980 
481 
826 
548 
7 
1981 
856 
1,978 
1,097 
6 
1982 
276 
643 
2,815 
44 
1983 
326 
784 
1,766 
23 
1984 
772 
2,037 
3,037 
15 
1985 
1,800 
4,722 
950 
2 
1986 
970 
2,136 
493 
2 
1987 
669 
1,400 
1,151 
8 
1988 
838 
2,074 
526 
3 
1989 
391 
767 
615 
8 
1990 
295 
676 
504 
7 
