FISHERY BULLETIN: VOL. 83, NO. 2 



In fact, the atretic condition of the ovary was a 

 more sensitive index of seasonal changes in the 

 reproductive rate among size classes of females 

 than was the incidence of spawning based on the 

 presence of postovulatory follicles. 



Atretic state 1 (<50% of yolked oocytes in the 

 alpha stage of atresia) was not useful for estimat- 

 ing atretic rates in an absolute sense since this 

 state appeared to persist in natural populations 

 for extended and probably variable periods. Some 

 spawning occurred among females classed in 

 atretic state 1, although the frequency of spawning 

 was less than half of that of females without 

 ovarian atresia. Batch fecundity might also be 

 reduced in females classed in atretic state 1, a 

 speculation worth further study. Atretic state 1 

 was a useful index of atretic rates during peak 

 spawning months. At such times it was the most 

 common atretic condition and detection of differ- 

 ences in atretic rates among length classes was 

 largely a function of the number of females in this 

 state. 



Atretic state 2 (5(y7f or more of yolked oocytes in 

 alpha atresia) persisted for about 9 d in the 

 laboratory, and judging by its low frequency in 

 field collections this state may have a similarly 

 short duration in natural populations. Females 

 with ovaries in this state rarely or never spawn, as 

 might be expected, since more than half of the 

 yolked oocytes are not viable. In addition, a short 

 duration of this state also might be expected on the 

 grounds that it seems maladaptive to prolong such 

 a threshold condition. For the above reasons 

 atretic state 2 seems to be the best absolute mea- 

 sure of the rates of ovary resorption in the popula- 

 tion and the only state that might provide an accu- 

 rate forecast of the end decline of reproduction in a 

 population. Unfortunately, accurate forecasts of 

 the end of spawning for a population can be made 

 only near the end of the spawning season. 



Atretic state 3 (no yolked oocytes with beta at- 

 resia present) identifies females in late post- 

 spawning condition. Such females cannot be sepa- 

 rated from immature females on the basis of gonad 

 weight or using gross anatomical criteria. This 

 state persisted for about 30 d in the laboratory, but 

 it may last much longer under natural conditions 

 while the numerous small oocytes are resorbed. 

 The laboratory data indicate that the duration of 

 this state could be increased if the definitions were 

 changed to include gamma + delta stages of atresia 

 which have a longer life in the ovary than the beta 

 stage. The laboratory data also indicated that 

 even gamma + delta stages of atresia would even- 



tually disappear from the ovary so that no signs of 

 previous spawning activity would exist in a re- 

 gressed ovary. It is doubtful that the duration of 

 atretic state 3 or any late postspawning state will 

 ever be accurately estimated because it is depen- 

 dent on too many environmental circumstances. 

 Nevertheless, this state is very useful in separat- 

 ing females in postspawning condition from 

 females with no previous reproductive history. 

 This is an essential distinction for estimating 

 spawning biomass (Stauffer and Picquelle foot- 

 note 3) and for determining the size or age at first 

 reproduction (Hunter and Macewicz 1980). 



Possibly the most important future application 

 of atretic classification of ovaries is for process 

 oriented sea work on the reproductive biology of 

 multiple spawning fish such as the northern an- 

 chovy. Such work does not require a large sample 

 as do estimates of reproductive characteristics for 

 an entire population. The reproductive state of an 

 individual female can be accurately defined by the 

 atretic criteria we have discussed, and the spawn- 

 ing state criteria described by Hunter and 

 Goldberg (1980). The reproductive characteristics 

 of a female can be related to its physiological state 

 (age, fat content, biochemical composition, and in- 

 stantaneous growth rate from otoliths or RNA/ 

 DNA ratios) and functional relationships estab- 

 lished between reproduction and the environ- 

 ment. In this way the factors controlling the 

 duration of the spawning season, and the total 

 fecundity during the season, can be identified 

 under natural conditions. 



Biological Implications 



Several important biological conclusions can be 

 drawn from this work. Only a few attempts have 

 been made to estimate the time needed for a folli- 

 cle to disappear by atresia in vertebrates and no 

 information exists for fishes (Byskov 1978). Our 

 focus was on atretic rates of all oocytes in the 

 ovary and not on an individual follicle; neverthe- 

 less, the striking speed with which all yolked oo- 

 cytes passed through the initial stages of atresia 

 indicate that the rate for individual follicles must 

 be high. Similar rates were observed in the guppy 

 by Lambert (1970a). In the guppy, alpha stage 

 atresia of yolked oocytes appears about 1 d after 

 parturition, and beta stage atresia appeared about 

 2 d after the first alpha stages were detected; beta 

 stages persisted for only 11 d. In the anchovy, the 

 average time for all yolked oocytes in the ovary to 

 pass through alpha atresia was 8.0 d and the 



134 



