Broodstock management, fecundity, egg quality and the timing of egg production in the rainbow trout (Oncorhynchus mykiss

Aquaculture, 100 (1992) 141-166 Elsevier Science Publishers B.V., Amsterdam

141

Broodstock management, fecundity, egg quality and the timing of egg production in the rainbow

trout (Oncarhynchus rnykiss)

%a11 Bromage, John Jones, Clive Randall, Mark Thrush, Briony Davies, John Springate, Jim Duston and Gavin Barker

institute of Aquacuhv, Univmity ofStiriing, Stirling FK9 4L.A UK

ABSTRACT

Bromage, N., Jones, J., Randall, C., Thrush, M., Davies, B., Springate, J., Duston, J. and Barker, G., 1992, Broodstock management, fecundity, egg quality and the timing of egg production in the rainbow trout ( Oncortrynchus my&s). Aquaculture, 100: 14 I- 166.

The full potential of rainbow trout hatcheries can be attained only if the hatcheries can provide producers with regular supplies of high quality eggs and fq every week of the year. This review as- sesses aspects of brookstock management central to determining total number of eggs produced, quality of eggs, and timing of maturation and spawning. The effectiveness of hatchery prcgrammes in supporting an expanding rainbow trout market is profoundly affected by the conditions under which the broodfish are maintained, their husbandry, and the stock selected for production. The paper ex- amines factors that can critically affect fecundity, egg production, egg quality, spawning time, and maintenance of egg supplies. The discussion also outlines methods of manipulating spawning and areas requiring further advancement uf knowledge if progress is to continue in optimising egg and fry production.

INTRODUCTION

The primary concern of any fish hatchery is to produce the maximum number of the highest quality eggs and fry from the available broodstock. This is particularly important in the farming of the rainbow trout because it is estimated that in excess of 3 billion eggs per annum are needed to support current world production of some 300 000 tonnes. As the number and quality of the eggs produced can be gTofoundly a.ffected by the conditions under which the broodstock are maintained, their husbandry, type of diet and ration and the stock or strain of broodfish selected by the hatchery, all these factors must be optimised if the full potential of hatcheries is to be realised and further growth in world production of trout is to be maintained.

0f parallel importance is an ability *;;o control the timing of reproduction of’

0044~8486/92/$05.00 0 1992 Elsevier Science Publishers B.V. All rights reserved.

142 N. BROMAGE ETAL.

the broodstock. This is becoming increasingly important because the major retail outlets for fish, and rainbow trout in particular (i.e. the chain multiple, freezer and supermarket stores), are demanding a continuity of supply of table fish of consistent size and quality every week of the year. This requirement can be met, only if hatcheries are able to provide production farms with regular supplies of eggs and fry for on-growing to market size.

The present review considers those aspects of broodstock management which are of importance in determining the number of eggs produced, their quality, and ihe timing of their supply in relation to the demands of the market place. Clearly, genetic selection programmes, biotechnology and aspects of reproductive physiology and endocrinology also have their parts to play in the management of trout broodstocks and optimisation of egg production. However, these subjects are considered by other authors elsewhere in this volume.

FECUNDITY AND EGG PRODUCTION

The number of ripe or mature eggs produced by salmonid broodfish is read- ily ascertained at full maturity by artificially stripping the eggs from the fish; this number is known as fecundity. Fecundity may be expressed in terms of the number of eggs produced per brood&h, when it is sometimes referred to as total or absolute fecundity, or more usually just as fecundity. Alternatively, fecundity may be expressed per unit body weight of post-stripped fish, when it is known as relative fecundity. Scientifically, there are objections to using relative fecundity because the number sf eggs produced for each unit increase in weight shows significant linear variation (see Bagenal, 1973, 1978; Brom- age et al,, 199Oa). However, it is a useful working index for the farmer because it allows egg production capability to be directly related to stocking levels, feeding rates, water supply, eMuent constraints, and the age and number of brood&h.

Estimates of fecundity of trout should only be made on fish when they are fully mature and ripe because only the eggs released at spawning represent the true reproductive capability of the broodfish. Determinations of fecundity made earlier in the cycle by macroscopic or microscopic examination of the ovaries of pre-spawning fish will be both optimistic and misleading because not all the eggs which are maturing in the ovary are ovulated and oviposited. Some eggs remain in the ovary and are subject to post-ovulatory resorption or atresia; others are resorbed before final maturation and ovulation occurs (Springate et al., 1985; Bromage and Cumaranatunga, 1987,1988); and lastly, varying numbers of eggs, although ovulated, are retained in the body cavity after artificial stripping. There are few records of the actual numbers of eggs which are retained after stripping although Briggs ( 1953) described reten- tions of approximately 10% in his review. Recently, work in this laboratory,

BROODSTOCK MANAGEMENT AND EGG PRODUCTION 143

in which actual egg counts were made on several hundred fish by sacrificing the broodfish and counting eggs remaining in the body cavity after artificial stripping, showed 1 O-20% of the estimated fecundity was retained, with higher percentages usually found in heavier broodfish (Jones and Bromage, unpublished). Generally, these retained eggs would be removed from the fish by a second stripping. However, few commercial farmers perform second strip- pings because the eggs produced are said to be of poorer quality and the eggs and broodfish are often damaged by the additional handling.

Two methods are used to estimate fecundity. The first method, based on the work of Von Bayer ( 1950), relies on measurements of mean egg diameter and the total volume of water-hardened eggs. This method is widely used by fish farmers and fishery scientists (Buss and McCreary, 1960; Satia et al., 1974; Baiz, 1978; Leitritz and Lewis, 1976) and provides estimates which are highly correlated with actual counts of eggs ( r= 0.998, PC 0.00 1) . The second method involves measurement of the total weight of eggs produced and the weight of a sub-sample of 100 eggs and thus, by proportion, an estimation of fecundity. This gravimetric method provides similar levels of accuracy to the volumetric one although it has the disadvantage of requiring an accurate mi- crobalance which may not be available on all farms.

Generally, when the fecundity of broodfish is described, some mention is made of egg size as this is thought by some workers to be an important deter- minant of egg quality (see later in this review). Sometimes egg size is described by egg diameter although more usually it is expressed -‘n terms of the number of eggs per unit volume or weight of eggs (e.g. eggs per litre ).

A further index of egg production, not normally used by farmers, is the total egg volume produced by each fish. This is a superior measure of egg produc- tivity because it takes into account both egg number and egg size and, hence, is more reflective of the reproductive properties of broodfish than either parameter on its own.

Fecundity, egg size and, in turn, total egg volume all increase with increasing fish size (Bromage and Cumaranatunga, 1988; Bromage et al., 1990a ). Typical regression equations for fecunditv and egg sias on post-stripped fish weight (Fig. 1) are as follows:

TF= 1347+ 1282 Wt (kg);P<O.OOl, v2=M%

OD=3.93+0.30 Wt (kg); P<O.OOl, r2=4:I%

where TF is total fecundity, OD is ova diameter and Wt is post-stripped weight obtained from data on 2468 females over a 0.5-6.0 kg weight range. Using length as a measure of size rather than wei#t gives similar levels of significance for the regressions. Log-log plots u~~ually improve the coefficients of determination of these regressions by about 5%.

Although both fecundity and ova diameter r.egressions are highly signifi-

144 N. BROMAGE ET AL.

8000

6000

0-3 0 1 2 3 4 5 6

WEIGHT (Kg)

Fig. 1. Regression of total fecundity (TF) and ova diameter (OD) on post-stripped fish weight.

cant, total fecundity on weight, with 67% of the data points covered by the

line of the regression, is far more strongly related to fish size than is egg size. In fact, with some groups of fish, specifically 2-year-old first spawners, egg size does not appear to be at all related to fish size. Possibly, factors other than fish size or nutrient supply determine egg size of first-spawning fish Multiple regression analysis, in which egg size is expressed as a function of both parental size and fecundity, greatly improves the strength of the relationship, in comparison to either parameter considered individually, suggesting that egg size is rather more dependent on the interaction of fish size and fecundity (Jones and Bromage, unpublished). Similar relationships for trout have been reported by other workers (Nicholls, 1958; Nomura, 1963; Bulk- ley, 1967; Kato, 1975; Springate and Bromage, 1984a; Bromage and Cumar- anatunga, 1988; Bromage et al., 1990a) although only Bromage et al. ( 1990a) fully validated the relationships statistically.

Regressions for total egg volume on post-stripped fish weight (Fig. 2 ) generally show even a stronger correlation with 75% or more of the data points described by the line of the regression. Consequently, it is recommended that this parameter be used when making detailed comparisons of egg producing capabilities of different broodfish,

As a consequence of the increase in egg size and the gradually diminishing rate of increase in fecundity with increasing fish size, relative fecundity de- creases as the fish get larger. This gives the appearance elf a compensation or “trade-off” between egg size and fecundity, thus maintaining a very high correlation between total egg volume and fish size in different sized fish (Roun- sefell, 1957; Gall, 1975; Springate and Bromage, 1984a; Springate et al., 1985; Bromage and Cumaranatunga, 1988; Bromage et al., 1990a). Clearhi, egg size, fecundity and fish. size comprise a complex of interrelated characters; change one and there are compensatory alterations in the others.

The strength of the relationship between fecundity and fish size makes it


2.50

t

4 LogTV a (ml)

2.25 l ’

Fig. 2. Log-log regression of total egg volume (log TV) on post-stripped fish weight (log Wt ). From Bromage and Cumaranatunga ( I988 1.

difficult to establish whether other factors are involved in the control of fecundity. This differentiation can only be achieved by partitioning the effects of fish size on fecundity or egg size using appropriate regression and covari- ante procedures (Snedecor and Cochran, 1980). Experimental protocols must also take account of the relatively high coefficients of variation of methods of fecundity estimation; this means that statistical significance is unlikely to be achieved unless experimental groups have n > 20.

One factor shown to influence the egg producing capability of trout is the stock or strain of broodfish. Recently, Bromage et al. ( 1990a), in a study of 12 commercial stocks of rainbow trout, have shown that the rates of increase of fecundity with increasing fish weight were similar for the different strains (i.e., the regressions all had similar slopes) but there were highly significant differences in the elevations of the regressions. Some stocks produced almost twice as many eggs as the least fecund even after differences in fish size were removed by covariance analyses ( Fig. 3 ) . Differences between stocks of trout of the same order of magnitude have also been reported by Nicholls ( 1958 ), Nomura ( 1963) and Bulkley ( 1967). In Bromage et al.‘s ( 1990a) study, regression of total egg volume on weight showed that all strains shared a com- mon slope but, again, there were highly significant differences in elevation with some strains producing up to a 55% greater volume of eggs. Although there were differences between strains in egg size, overall these were less im- pressive, with only a 10% difference between strains producing the largest and smallest eggs. 6ne strain, which is reputed to have been selected for increased fecundity, was especially interesting because egg size remained constant throughout the weight range of the broodfish.

N. BROMAGE ET AL.

6

Fig. 31 Log-log regression of total fecundity (log TF) on postcstripped fish weight (log Wt ) of 12 strains of rainbow trout. From Bromage et al. ( 1990a).

There are important nutritiona”l effects on fecundity. A number of studies have reported that fecundity is increased by feeding high energy diets or larger daily rations (Phillips et al., 1964; Smith et al., 1979; Harris and Griess, 1978; Orr et al,, 1982; Springate et al., 1985). Periods of diet restriction or starvation have variously been shown to reduce (Vladykov, 1956; Scott, 1962; Baiz, 1978) or be without effect on fecundity (Kato, 1975; Ridelman et al., 1984). None of these studies, however, used covariance analyses to adjust for differences in body size expected from reductions in diet or starvation. Thus, direct dietary or ration effects on fecundity were not proven. By contrast, Bagenal ( 1969) and Roley ( 1983), each using two feeding levels, were able to dem- onstrate dietary intake effects on fecundity that were independent of the fish size-fecundity relationship.

These studies were recently extended IQ Jones and Bromage ( 1987) who investigated the effects of feeding large grumps of rainbow trout at rates of 0.4%, 0.75%, 1 .O%, 1.2% or 1.5% of fish body weight per day for a year before their first spawning as 2-year-old fish. Broodfish on 0.4% and 0.75% rations produced significantly fewer eggs than those fed at the three higher rates even after due allowance was made for the large differences in fish size (Fig. 4). Total egg volume also showed parallel reductions on the two lower rations. Although the mean egg size was larger for the fish on the higher rations, none of the regressions of egg size on fish weight was sigtiificant, From a management viewpoint, there would appear to be no advantage of feeding in excess


I

4 -0.4 -0.2 0.0 0.2 0.4 0.6

LOG POST4TRIPPED WEIGHT (Kg)

Fig. 4. Log-log regressions of total fecundity on post-stripped fish weight of five groups of rainbow trout fed on different daily rations (% body weight per day) as follows: group 1,0.4%; 2, 0.75% ; 3, 1%; 4, 1.2%; and 5, 1.5%. From Jones and Bromage ( 1987).

of 1 .O% of fish body weight per day. An important finding was that 16% fewer fish spawned on the 0.4% and 0.75% diets. This confirms other studies by Bagenal ( 1969) and Springate et al. ( 1985).

Most of these feeding studies employed the same levels of ration throughout the year. Intuitively, one expects that fish in the wild would have a seasonally-variable requirement, possibly with higher demands during the periods of most rapid egg growth. Surprisingly, only Ridelman et al. ( 1984) have investigated the possible effects of seasonal alterations in ration. These workers reported that starving fish for 1.5 months before spawning was without effect on fecundity and egg size. However, one might question how much food was being eaten by the fed group as both the starved and control fish were of similar weight at spawning.

Recently, further investigations of the effects of seasonal alterations in ration have been carried out in our laboratory (Jones and Bromage, unpublished). Significant changes in fecundity (Fig. 5) and total egg volume, and in the relationship of these parameters to fish weight resulted from changes between high ( 1 .O% body weight/day) and low rations (0.4%) at different times of the year (Fig. 6). This study, which was carried out on 1 -year-old fish over the year before their first spawning as Zyear-olds, also utilized CO- variance techniques to account for any fish size-related differences in fecundity.

Reducing the ration over the last 3 months before spawning (Group 6, Fig. 4) had no detrimental effects on fecundity. In fact, relative fecundity was significantly increased as a result of the lower rates of growth in the 3 months leading up to spawning. Feeding at a higher rate for the period 5-9 months into the cycle (Group 3 ) resulted in significantly higher fecundity and total egg volume than for fish fed at lower rations over the same period ( 6rm.w 4).


-0.3 -0.2 -0.1 0.2 0.3

8 LOG POST~STRI& WE& (Kg) 4

Fig. 5. Log-log regression of total fecundity on post-stripped fish weight of groups of rainbow trout each fed for different periods of time on high ( l.O”h body weight/day) and low (0.4% body weight/day) rations; see Fig. 6 for protocol of experiment. From Jones and Bromage (unpublished),

GlOUD 1 I Gram 2

Group 3 Group 4

I I

1 I

Group 5 GroUD6

Fig. 6. Protocol of experiment to investigate the effects of different periods of high and low ration on fecundity, egg spawning. From Jones and Bromage (unpublished).

feeding 2-year-old rainbow trout for size, total egg volume and % of fish

Increases in fecundity were also seen for fish fed high rations for the first 4 months of the cycle (Group 2) whereas fish receiving low rations over this period (Group 1) seemed to have somewhat limited fecundity despite good somatic growth over the succeeding 8 months before spawning. Collectively, these results suggest that fecundity is established quite early in the annual cycle.

There were no significant differences in egg size for the fish under the different feeding regimes although those fish receiving high ration throughout the year (Group B) or during the period 5-9 months into the cycle (Group 3) produced batches of the largest and smallest eggs, respectively.

In addition to the effects on fecundity, there also was a clear relationship between ration and the percentage of fish which spawned. Those groups fed


high rations for the first 4 months of the cycle (Groups B, 2,4 and 6) had the highest percentage of spawned fish; of these, groups B, 2 and 6 had some of the highest fecundities. By contrast, fish maintained on low rations (Groups C, 1 and 5 ) exhibited 3O-35% reductions in number of maturing fish. Al- though the fish on the lower rations that did spawn had higher relative fecundities, overall, their smaller size and reduced levels of maturity would mean that a farm using such feeding regimes would have to maintain three times as many brood&h to produce the same number of eggs (Table 1). Furthermore, it would appear that the most productive group of broodfish was Group 6 and not those maintained on high ration throughout. This has important implications in the management of broodstock. I_ _

From the foregoing experiments emploqiing different levels of feeding to induce changes in fecundity and body size, it would appear that although fecundity bears a relatively constant relationship to fish size, broodfish that achieve a major proportion of the yearly growth during the last third of the reproductive cycle have lower fecundities than we would expect from their weight. This suggests that a finite range of potential fecundity is established earlier in the reproductive cycle.. Subsequent changes in fecundity do occur but this constitutes only a fine tuning of the number of eggs that ultimately will be produced at spawning. Unfortunately, at present there are few data which enable us to understand how fecundity is controlled and the important environmental and physiological determinants. We also have limited information on the dynamics of these processes.

Changes in fecundity could be achieved by modification of the rate of recruitment of previtellogenic oocytes (stages I,2 and 3: Bromage and Cumar- anatunga, 1987, 1988, Fig. 7) and in turn cortical alveoli oocytes (stage 4)

TABLE I

Effects of varying patterns of feeding high and low rations rsee Fig. 6 for protocol) on % of fish spawning, mean fecundity and mean weight of female rainlrrrtq trout. -A,k ix!uded are derived estimates of numbers and biomass of fish on the different feeding regimes required trs produce 1 million t2ggS.

Group Spawning VW

Mean Fecundity Mean weight (kg)

No. of fish’ Total fish biomass’ (kg)

B 68.2 3036 1.571 483 C 34.5 1864 0.775 1555 1 48.2 2562 1.321 810 2 63.6 2355 0.903 668 3 47.2 2693 1.047 785 4 68.2 2268 1.305 646 5 40.9 1865 1.081 1311 6 70.0 3060 1.163 467

.~ .~ ‘Number and biomass kf fish reqlrirpd ?C k: xkxc : iiliiliori eggs

759 1205 1070 603 822 843

1417 543

N. BROMAGE ET AL.

o- QOGONILI; 1, 2a. 2b, 2C, 3. 4% 4b- PREVIlCLLO~ENlC OotVTES; 5, 6, 7- VlTELLWXNlC OOCYTES; m- POST-OVUUTORI FOLLlCFEm

Fig. 7. Drawings of sections of oocytes of rainbow trout illustrating the different stages of the oocyte development cycle ( x 10 magnification). From Cumaranatunga and Bromage (unpublished ).

into vitellogenesis. Certainly sufficient numbers of these oocytes are present in the ovaries of all fish in the year preceeding each spawning to provide the cohorts of 2000 oocytes per kg of body weight that complete exogenous vitellogenesis to become fully mature and ovulated. Oogonia, although present throughout the cycle, are found in greater numbers immediately after spawning. Their transformation to previtellogenic oocytes at this time would coin-

BROODS-OCK MANAGEMENT AND EGG PRODUCTION 151

tide with the suggestions made above that fecundity is determined early in the year-long cycle of development preceding spawning.

The number of maturing oocytes can also be modified by atresia (Cumar- anatunga et al., 1985; Springate et al., 1985; Bromage and Cumaranatunga, 1987, 1988). Atresia is said to occur at all stages of oocyte development although it has only been clearly shown to affect stages 4,5,6,7 and unovufated oocytes (Bromage and Cumaranztunga, 1987, 1988 ) . Generally, atresia only occurs amongst stage 4 oocytes in ovaries post-ovulation or in ovaries of fish that have been starved or subjected to hormonal treatment (Bromage and Cumaranatunga, 1987, 1988 ). Atresia, however, is found at varying levels in all vitellogenic ovaries, appearing as soon as oocytes begin sequestering exogenous yolk. Following periods of reduced feeding, up to 22% of the total number of vitellogenic oocytes may become atretic (Springate et al., 1985). Starvation may produce 100% atresia amongst vitellogenic oocytes (Bromage and Cumaranatunga, 1987, 1988). This may be the reason why some salmonids miss a spawning if subjected to poor conditions. However, under normal circumstances less than 10% of the vitellogenic oocytes become atretic ( Hen- derson, 1963; Springate et al., 1985; Bromage and Cumaranatunga, 1987, 1988 ) and this may provide the “fine tuning” of fecundity which clearly occurs during the later stages of the reproductive cycle.

EGG QUALITY

Good quality eggs are usually defined as those which exhibit low levels of mortality at fertilization, eying, hatch and first-feeding and those which produce the fastest-growing and healthiest fry and older fish. There is, however, no general agreement as to what levels of mortality constitute a good quality egg nor is there any understanding of the factor or factors ira the egg or the broodfish responsible for differing standards of quality. What is clear is that there is considerable variation in egg quality even in eggs produced by different individuals of the same stock maintained in the same tank under appar- ently identical conditions.

Almost 40 years ago Briggs ( 1953), reviewing other work in a number of US hatcheries, reported that egg losses to eying averaged 18-19%. Similar losses to eying were also reported by Springate and Bromage ( 1983, 1984b,c ) and Bromage and Cumaranatunga ( 1988), some 30 years later, in extensive surveys of rainbow trout hatcheries in the UK (Fig. 8 ) . This was a somewhat surprising finding given the general improvements in farming methods that have occurred. These authors also described mean fertilization and hatching rates of 90% and 70%, respectively, with survivals of fry up to 4 months of age of only 35-40% (Fig. 9 ). Within these mean survival rates, however, there was considerable range in individual values, with some batches having survivals up to 4-month fed-fry in excess of 85%, whilst in others, mortalities of

Rainbow trout l

% Survival (to winal

i

Farm p Q R ST

Brown trout

u w Overall I x meen

N. BROMAGE ET AL.

Fig. 8. Survival (mean % 31 s.e.m. ) to the eyed stage of eggs of differen? ages of brown trout ( x )

and rainbow trout (P-W) from seven commercial farms. From Springate and Bromage (unpublished ) and Bromage and Cumaranatunga ( 1988 ).

Fig. 9. Survival rates (mean %k s.d. ) after fertilization and to the eyed, hatch and swim-up stages and as fry up to 4 g of rainbow trout maintained under commercial conditions.

100% were found at fertilization. There are similar reports of 100% mortalities (“blanks”) by other workers (Craik and Harvey, 1984; Ridelman et al., 1984) and in other fish (Carillo et al., 1989).

Most of the data relating to egg quality for trout are complicated by the use of different experimental conditions and treatment of broodfish. Generally, control groups from other experimental protocols have to be examined to provide informatidn on survival. Thus, Lincoln and Scott ( 1983) and Happe et al. ( 1988)) in two studies on triploidy, reported hatching rates of 59% and 70-90%, respectively, for the two control groups of eggs. By contrast, Craik


and Harvey ( 1984) described survivals of only 28% up to the same stage,

although their study was carried out on younger broodfish and included a number of zero survivals. In a selection programme to improve the growth rate of commercial stocks of rainbow trout conducted over a 7-year period, JSincaid et al. ( 1977 ) recorded mean survivals of 40% up to 147 days post- fertilization. Similar levels of survival (45%) were also reported by Springate et al. ( 1985 ) in a study of the effects of broodfish ration on egg quality. In other nutritional investigations, Smith et al. ( 1979) and Roley ( 1983) found eying rates of 79% and 59%, respectively, in both control and experimental groups and Ridelman et al. ( 1984) found survival to hatch of 43% for eggs from control fish and those which had been starved for lc5 months before spawning. In a study of the relationship between chemical composition and survival of rainbow trout eggs, Hirao et al. ( 1955 ) recorded a mean eying rate of 73%. Recently, Bromage (unpublished) in a further survey of survival amongst 10 million eggs in commercial hatcheries found eying rates ranging from 80% to 90% in different batches.

Although mean survival to eying in these various studies would appear to approximate to 80%, it is evident that many individual batches of eggs from individual fish are of much poorer quality. Possibly, it is on these poorer quality eggs or “blanks” that we should focus our attention. Improvements here would be of considerable economic value to hatcheries. Many factors have been implicated as possible determinants of egg quality, including the nutrition of the brood&h, the chemical composition of ?he egg, the size of the egg, and over-ripening of the eggs, a change which is determined primarily by the timing of stripping of the broodfish. Up to present, however, aside from over- ripening, there are few consistent data to explain variation in egg quality.

In general, feeding broodfish different levels of ration, although affecting fecundity and possibly egg size, does not appear to have any effect on egg quality (Roley, 1983; Ridelman et al., 1984; Springate et al., 1985; Knox et al., 1988). By contrast, investigations of the effects of experimental alterations in diet composition are conflicting. Smith et al. ( 1979 ) and Satia ( 1973 )

both showed dif’Cerences in survival when different formulations were fed but the diets were so different that proper conclusions could not be drawn from these data. Takeuchi et al. ( 198 1) reported higher hatch rates for eggs from trout maintained on a low (36%) protein diet, whereas Roley ( 1983) found that the eggs produced by broodfish fed on a 47% protein diet had higher survival than those produced by fish on either 27% or 37% protein diets. Clearly, further work is required in this area.

Regarding lipid requirements, there is little information on the levels of fat and types of EFA (essential fatty acid) required in the diets of trout broodfish and no published study of the effects of differing fat levels on egg quality in this species. We have shown that fecundity and the costs of egg production are optimised by using diets with 7- !! 2% gross fat (Jones and Bromage, un-


published), but were unable to show any differential effects of 7%, 12%, 18%

or 25% fat on egg quality. In his review of salmonid broodstock nutrition, Hardy ( 1983 ) concluded that omega-3 fatty acids were essential for trout broodfish, with some indications that longer chain forms may promote egg quality. Watanabe ( 1985 ) suggests that rainbow trout broodfish may have some requirement for omega-6 lipids.

Work on minerals and vitamins has tended to concentrate on chemical analyses of the eggs with the aim of correlating differing levels of the various constituents with variation in egg quality. Although there are differences in egg composition, particularly of the trace metals (Craik and Harvey, 1984; Springate et al., 1985), most of this variation is only weakly, if at all, correlated with patterns of survival. Furthermore, it would appear that egg quality is reduced only, when trace elements or vitamins are completely eliminated from broodfish diets and as a consequence, are markedly deficient or totally absent from the eggs (Takeuchi et al., 198 1; Hardy, 1983; Sandnes et al., 1984). With modern diets, methods of quality control and current feed prac- tice on farms of mixing diets from different manufacturers, it is unlikely that such deficiencies would be found at commercial hatcheries. Hardy ( 1983 ) concluded that increasing dietary vitamin and trace elements from deficient to adequate levels can influence the levels of materials in the eggs and improve their quality. However, further supplementation does not seem to offer any further advantage and may, in the case of the fat-soluble vitamins, ac- tually reduce egg survival, possibly by increasing the competition by individual vitamins for intestinal transport and absorption. Excessive vitamin supplementation has been suggested as a possible contributory factor to the development of “fry anaemia” or “rainbow trout fry syndrome”, a disorder of unknown aetiology which is-producing fry losses of up to 50% in some European hatcheries.

Amongst other factors implicated as likely determinants of egg quality, egg size has been the subject of much controversy. Some workers suggest that smaller eggs experience increased mortality (Small, 1979; Pitman, 1979) whereas other suggest that size does not have any effect on egg quality (Glebe et al., 1979; Kate and Kamler, 1983; Thorpe et al., 1984; Springate and Brom- age, 1985 ). Contributing reasons fc4r the confusion include differences in age and size of parental fish, varying culture conditions and, probably of greatest importance, uncontrolled variation in ripeness of the eggs (see later). Pro- vided eggs are stripped 4- 10 days after ovulation at 10 O C, and given appropriate husbandry and water supplies, small eggs produce similar rates of fertilization to larger ones (Fig. 10) (Springate and Bromage, 1985; Bromage and Cumaranatunga, 1988 ) . Also, there is no relationship of egg size to morn Ulities at eying, hatch, swim-up or as older fish (Springate and Bromage, 1985; Bromage and Cumarantunga, 1988 ) . Larger eggs do, however, produce larger first-feeding fry (Pitman, 1979; Springate et al., 1985; Springate and Brom-

BROODnOCK MANAGEMENT AND EGG PRODUCTION 155

0 0 (mm)

Fig. 10. Survival rates (% rate of fertilization: %F) of different sizes of rainbow trout eggs (ova diameter: CID). From Springate and Bromage (unpublished) and Bromage and Cumarana- tunga (1988).

age, 1985 ). Given such differences in absolute size and hence mouth gape, it is possible that smaller fry might suffer higher mortalities if they did not re- ceive enough food or were fed on a pellet or crumble that was too large. Smaller eggs may also be subject to relatively higher disease risks than larger eggs as a result of their larger surface area to volume ratio (Bromage and Cumarana- tunga, 1988 ); there is evidence that the incidence of increased numbers of bacteria on the outside of trout eggs is correlated with reductions in hatch rate (Barker et al., 1989 ) c

Although larger eggs produce significantly bigger hatched fry, the weight of evidence in the literature suggests that this size advantage is soon masked by other environmental determinants of growth. Springate and Bromage ( J. 985) showed that the differences in size of fry resulting from differences in egg size were not apparent statistically, as early as 4 weeks after first-feeding. Kincaid ( 1972 ) and Springate et al. ( 1985 ) also reported disappearances in this size advantage but not until 150 and 130 days of growth, respectively. Springate and Bromage ( 1985) and Bromage and Cumaranatunga ( 1988) also showed that the fry hatched from different sizes of egg had similar specific growth rates and their potential for growth was equivalent. Thus, it would appx that under good hatchery conditions differential egg size is not a primari de- terminant of egg and fry quality.

At present rhe only factor which has been shown clearly to affect egg quality in the rainbow trout is the time of stripping of fish in relation to ovul,&on (Nomura et al., 1974; Sakai et al., 1975; Escaffre and Billard, 1979; Springate et al., 1984). Under farm conditions, the eggs of rainbow trout are ovulated but not oviposited; they remain in the body cavity until they are artjficially stripped from the fish. During the period of retention in the body cal-ity, the eggs undergo a ripening process. Eggs stripped between 4 and 10 dz,ys after ovulation at 10 O C consistently achieve high rates of fertilization (Springate et al., 1984). There are modest reductions in fertility if eggs are stripped immediately after ovulation, possibly because some of the eggs may be forced


out by the manual stripping or are slightly immature or ‘under-ripe’. At lO- 15 days after ovulation at 10” C, and after shorter or longer times at higher and lower temperatures respectively, significant reductions in egg quality occur. After 20 days, only a few eggs .are capable of being fertilized and all of these usually die during incubation.

Eggs with moderate or low rates of fertilization invariably experience similar low rates of survival to eying, hatch and swim-up. Therefore, one might conclude that the fertilization rate, measured either after 10- 12 h or 7 days at 10” C, is a very good predictor of subsequent performance (Springate and Bromage, 1984b, 1985 ). Hence, there is little point in maintaining batches of eggs with poor fertility as there will be further high mortalities at each of the developmental stages of eyisg, ha.tch and swim-up.

TIMING 06 :IPAWNING AND THE SUPPLY OF EGGS

Under ambient conditions, different stocks of rainbow trout are said to spawn ir every month of the year (Bromage and Cumaranatungzt, 1988). In hatcheries, however, the majority spawn during the late autumn and early winter months, usually with each individual broodstock producing eggs over a 6-8 week period. This seasonality of spawning imposes considerable constraints on trout farming because the consequent restrictions on the supply of eggs and fry make it difficult for on-growing farms to maintain a continuity of production of table-size fish throughout the year. Ideally, hatcheries should artificially control the spawning times of their broodfish so that batches of eggs and fry might be produced all year round.

Hormonal treatments involving injections of LHRHa, HCG or pituitary extracts have been used to &Vance spawning, but only by 2-3 weeks; only the final stages of oocyte maturation can be accelerated if problems of poor egg quality are to be avoided. Such advancements do little to improve egg supplies although they may be helpful in synchronizing the spawning of groups of broodfish. Recent developments using implants of LHRHa may prove more effective (Crim et al., 1986 ).

Much more pronounced modifications of spawning time are possible using manipulations of photoperiod. Advances and delays of 4 months or more have been achieved (Bromage et al., 1984; BrcrTage and Duston, 1986 ) primarily because the treatment of broodfish can be initiated at the start of the year- long reproductive cycle. Methods of photoperiod control also have the advantage of being cheap and simple to install on commercial farms and, unlike hormonal treatments, do not involve handling the broodfish.

Photoperiodic manipulation is effective because under natural conditions the seasonally-changing pattern of daylength provides proximate cues that coordinate the timing of gonadal recrudescence, oocyte maturation and spawning with changing season. By altering the rate of change of daylength of


the yearly light cycle, maturation can be advanced or delayed for commercial advantage in a wide range of fish species including the rainbow trout (see reviews Bromage et al., 1984,1990b,c; Bromage and Duston, 1986; Bromage and Cumaranatunga, 1988). Many data also exist to show that the increasing and decreasing daylength components of natural and modified seasonal light cycles can be replaced by combinations of constant long and short days (Whitehead and Bromage, 1980; Bromage et al., 1982; Bromage and Duston, 1986). Thus, long days early in the reproductive cycle and short days at any time in the 3-4 months before the summer solstice advance rkiaturation, whereas short days during the first few months of the cycle or long days after the summer solstice delay gonadal development (Fig. 11) . The ability of trout to respond to photoperiod regimes comprising daylengths of constant length is particularly helpful because such regimens are much easier to implement on commercial farms than seasonally-changing ones.

The response of rainbow trout to photoperiods of constant length has led to hypotheses that daylength of specific duration imposes direct inductive influences on the timing of reproduction, in an analogous way to that reported for some other vertebrates (Follett, 1984). Thus, Breton and Billard ( 1977 ) considered trout to be “short day animals” because the final stages of maturation and spawning occur under short or decreasing daylength. Con- trastingly, trout were described as “long day animals” by Bromage et al. ( 1984) and Scott and Sumpter ( 1983) who proposed that the primary inductive event occurred at the beginning of the reproductive cycle when daylengths were lengthening.

There are a number of difficulties with these interpretations. Firstly, the response of trout to putative critical daylengths is very slow, taking several months to develop, and also varies depending on the timing of the photoper-

__~____ ADVANCE’

Fig. 11. Summary of effects on spawning time of exposure of rainbow trout to long and short daylengths at differen, ._ __ + +iI*res of the reproductive cycle. From Bromage and Duston ( 1986).


18 Group D

6

6) , b t , , , , , JFMAMJ J A S

Time (months)

Fig. 12. Observed and expected spawning times of groups of rainbow trout maintained on a series of acceleratory ‘long-to-short’ photoperiods (Groups A-D). Vertical axis units are h light day-l; months of the year on the horizontal axis. From Duston and Bromage (unpublished).

iodic cue (Fig. 12). By contrast, exposure to a critical daylength of some higher vertebrates induces an immediate response (Follett, 1984 ) . It is also clear, at least as far as trout are concerned, that any daylength can be considered as long or short provided they are preceded by daylengths that are shorter or longer, respectively. Thus, spawning has been shown to be advanced by LD 10: 14 ( 10 h light and 14 h dark each day) provided it is followed by LD 6: 18. There is little difference between this effect and the responses achieved following exposure of fish to LD 14 : 10 and th.en LD 10: 14 or LD 18 : 6 and then LD 14: 10 (Bromage and Duston, 1986; Duston and Bromage, 1987; Randall et al., 1987; Bromage et al., 1990b,c). Thus, photoperiodic history and the direction of change of photoperiod are of more importance than critical daylength in timing reproduction in the rainbow trout.

A further argument against a direct inductive effect of photoperiod is that spawning of trout occurs under a range of different constant daylengths from LD 6: 18 to ED 22 : 2 (Whitehead et al., 1978; Bromage and Duston, 1986; Davies, Randall and Bromage, unpublished). Spawning also continues under DD (constant darkness) and LL (constant light) (Bromage et al., 1984; Bromage and Duston, 1986; Duston and Bromage, 1986; Bromage and Dus- ton, unpublished). Collectively, these results are strongly suggestive of an internal or endogenous rhythm.

Firm acceptance of a rhythm as endogenous requires that its periodicity be followed over more than one and ideally multiple cycles, whilst maintaining a constancy of all conditions (see Bromage and Duston, 1986; Duston and Bromage, 1986, 1987, 1988, 199 1 for more detaiBed discussions). Such experiments have now been completed for maturation cycles of rainbow trout (Duston and Bromage, 1986,199 1); fish maintained under constant 6L: 18D, feeding rate, and water temperature and quality over a 5-year period spawned at intervals which approximated but were significantly different from a year


- ym’i v-aYcar 2 l 1Yepr3 f_Ytiu--ytpIQ~Yeu5-

Fig. 13. Spawning times of female rainbow trout maintained for 5 years under conditions of constant short days (LD 6: IS), temperature ( lO”C), feed rate (O-5’% body weight per day) and water quality. From Duston and Bromage ( 1986, ! 99 i j.

(Fig. 13 ) . This clearly establishes the presence and involvement of an endogenous rhythm or “clock” in the control of reproduction of rainbow trout. Un- der constant conditions this rhythm “free runs” with a periodicity of about a year, that is, it is circannual.

One would expect such a rhythm to respond to modified seasonal and constant light regimes by undergoing corrective advances or delays in its phasing, depending on whether the daylength to which the fish is exposed is perceived as running ahead or behind that of the internal clock. In this way, long days early in the year, that is, before the longer daylengths of the summer solstice, would produce an advance of the internal clock and in due course an earlier spawning, whereas long days later than the summer solstice would produce a phase delay of the rhythm and a delay of spawning. Fig. 14, taken from Brom-

(slight) (maximum)

Fig. 14. Schematic representation of the ‘phase relations* of the endogenous rhythm controlling rate of maturation and spawning time, induced by long day lengths administered at different times of the annual seasonal light cycle. Exposure of fish to long days in the period December- March results in a maximum phase advance. From Bromage et al. ( 199Pc).


age et al. ( 199Qc), gives a schematic view of the suggested phase response relationships of spawning time to treatment with long days or constant light; the seasonally-changing ambient light cycle is included to aid comparison. It should be mentioned that this model is compatible with all the results that have been reported for long day and constant light (LL) treatments. It also explains why 2 months of LL, in an otherwise ambient photoperiod, administered just after spawning or after the summer solstice, produced 4-7-month advancements and delays of spawning time, respectively (Figs. 15 and 16)

(Randall and Bromage, 1992 ) . By contrast, exposure of fish to short days would be expected to elicit quite

different responses from an endogenous rhythm. Short days early in the reproductive cycle result in a delay in maturation, probably because the internal rhythm or clock would have been thought to be running ahead of the perceived daylength (i.e. the short days of winter) and there would have been a delay in the phasing of the rhythm. Later in the spring, exposure of fish to the same short days produces an advance in spawning, presumably because this

No. p spawnmg

100

80

60

40

20

Fig, 15. Effects of 2 months of continuous light (LL) in an otherwise natural light cycle on the timing of spawning of female rainbow trout. Vertical axis units are numbers of female fish spawning: months of the year on the horizontal axis.

60

r

L-L-J

Fig. 16. Effect of 2 months of continuous light (LL) in an otherwise natural light cycle on the timing of spawning of female rainbow trout. Vertical axis units are percentages of female fish spawning; horizontal axis units are months of the year.

BROOD!XOCK MANAGEMENT AND EGG PRODUCTION 161

daylength would have been perceived to be the decreasing photoperiod of autumn and there would have been a corresponding phase advance in the rhythm controlling maturation. These changes in phase in response to short days are summarised schematically in Fig. 17 (Bromage et al., 199Oc).

Temperature is also known to affect reproductive development and the timing of spawning of rainbow trout (Morrison and Smith, 1986; Bromage and Cumaranatunga, 1988; Nakari et al., 1988) although there is little evidence, at least as far as salmonids are concerned, that the effect is due to temperature acting as a proximate cue. This view is supported by our recent finding that rainbow trout maintained on a river water supply, with temperature varying seasonally from 4 O C in February to 16.5 O C in July, experienced similar advances in spawning time in response to a stimulatory photoperiod, to a parallel group of fish maintained on the same light but receiving constant 8.5 “C borehole water (Davies and Bromage, unpublished ) . Probably, many of the reported effects of temperature are the result of direct influences on the ovary. Cultured oocytes certainly sequester yolk at slower rate3 when temperatures are reduced (Tyler et al., 1387 ). Whether such reductions in yolk se- questration also occur in vivo and whether this effect could account for observed delays in spawning time said to result from maintaining fish at temperatures of l-2°C (Morrison and S-mith, 1986; Nakari et al., 1988; Bromage and Cumaranatunga, 1988 ) is not clear.

The finding that the rate of maturati’on anid the timing of spawning are dependent on the entrainment of an endoge ous circannual rhythm by photoperiod change has been an important ‘stepforward’ in our understanding of the control of reproduction of the trout and possibly many other fish. Jt has

Phase advance (slight) (maximum)

Fig. 17. Schematic representation of the ‘phase relations’ of the endogenous rhythm controlling the rate of maturation and spawning time, induced by short day lengths administered at different times of the annual seasonal light cycle. Exposure of fish to short days in the period Decem- ber-March results in a phase delay of the rhythm. From Bromage et al. ( 1990~).


L I

I J

c - -I L---L_ 1 I I I I 1 I

May JUIY Sept Nov Jan Mar May July

Fig. 18. Photoperiodic manipulation of times of spawning of four different strains of rainbow trout (A-13) to achieve spawning fish in every month of the year. NB: some times are only achieved after 2 or more years of photoperiodic manipulation.

enabled us to understand why the same daylength can have quite different effects on maturation at different times of the year, thus allowing introduc- tions of broodstock into photoperiodic control to be more easily and effec- tively made.

When developing procedures for the all-year-round production of eggs, it is important to retain in the population any variations in spawning time, whether it is for different individuals of the same stock or between genetically-distinct stocks. The use of early and late spawning stocks makes it a lot easier to achieve significant advances and delays of egg-production capability by photoperiodic manipulation. By using strains with different natural spawning times in combination with advancing and delayibg photoperiods and out-of-season yearly seasonal light cycles or constant light regimes which induce circannual “free-running’ spawning periodicities (Fig. 18 ) , it is relatively easy to provide the all-year-round production of eggs and fry that the industry requires.

For the future, it is expected that the induction of the phase changes in the endogenous rhyttin by exposing fish to short periods of short, long or continuous daylength in an otherwise ambient photoperiod will be more widely ex- ploited by commercial farms. It has the advantage of not requiring full black- out and hence would not tie-up facilities and could be used at cages or other difficult sites. It also means that much larger numbers of broodfish might be maintained under photoperiodic control.

CONCLUSIONS

Egg-production capability remains one of the single most important determinants of successful trout production. Techniques being used for trout

BROODSTOCK MANAGEMEPIT AND EGG PRODUCTION 163

promise to be instructive for the development of methods of broodstock management for other species. At present we are able to control the spawning time of trout at will and produce modest improvements in fecundity. However, further advances in our understanding of the underlying determinants of egg quality and fecundity need to be made if the industry is to continue to flourish.

ACKNOWLEDGEMENTS

Portions of this work were supported by grants to N.B. from MAW, NERC, BP Nutrition (UK) Ltd and a number of commercial farms.

REFERENCES

Bagenal, T.B., 1969. The relationship between food supply and fecundity in brown trout Salmo frulfa. J. Fish. Biol., 1: 167-182.

Bagenal, T.B., 1973. Fish fecundity and its relations with stock and recruitment. Rapp. P.-V. Reun. Cons. Int. Explor. Mer, 164: 186- 198.

Bagenal, T.B., 1978. Aspects of fish fecundity. In: S.D. Gerking (Editor) ,Ecology of Freshwater Fish Production. Blackwell Scientific, Oxford, pp. 75 10 1.

Bait, M., 1978. Fecundity of reared rainbow trout Safmo gairdneri. Ecologia, 3: 57-64. Barker, G., Smith, L IT. and Bromage, N., 1989. The bacterial flora of rainbow trout and brown

trout eggs and its relationship to developmental success. J. Fish Dis., 12: 28 l-293. Breton, B. and Billard, R., 1977. Effect of photoperiod and temperature on plasma gonadotro-

pin and spermatogenesis in the rainbow trout. Ann. Biol. Anim. Biochim. Biophys., 17: 33 l- 340.

Briggs, J., 1953. The behaviour and reproduction of salmonid fishes in a small coastal stream. State of Calif. Dept. Fish. Game. Mar. Fish. Branch. Fish. Bull., No. 94, pp. 3-6 1.

Bromage, N. and Cumaranatunga, P.R.C., 1987. Oocyte development in the rainbow trout with special reference to vitellogenesis and atresia. In: D. Idler, L. Crim and J. Walsh (Editor), Reproductive Phy’siology of Fish. Marine Science Lab,, St Johns, Newfoundland, p. 194.

Bromage, N.R. and Cumaranatunga, P.R,C,, 1988. Egg production in the rainbow trout. In: R.J. Roberts and J.F. Muir (Editors), Recent Advances in Aquaculture, Vol. 3. Groom Helm, London, pp. 43-l 38.

Bromage, N. and Duston, J., 1986. The control of spawning in the rainbow trout using photoperiod techniques. Rep. Inst. Fresh. Res. Drottningholm, 63: 26-35.

Bromage, N., Whitehead, C., Elliot, J., Breton, B. and Matty, A., 1982. Investigations into the importance of daylength on the photoperiod control of reproduction in the female rainbow trout. In: C. Richter and H.Th. Goos (Editors), Reproductive Physiology of Fish. Pudoc, Wageningen, pp. 233-236.

Bromage, N.R., Elliot, J.A., Springate, J.R.C. and Whitehead, C., 1984. The effects of constant photoperiods on the timing of spawning in the rainbow trout. Aquaculture, 43: 2 13-223.

Bromage, N., Hardiman, P., Jones, J,, Springate, J. and Bye, V., 1990a. Fecundity, egg qizp and total egg volume differences in 12 stocks of rainbow trout. Aquacult. Fish. Manage., 2 1: 269- 284.

Bromage, N., Duston, J., Randall, C:. , Brook, A., Thrush, M., Carrillo, M. and Zanuy, S., 199Ob. Photoperiodic control of teleost r;e;production. In: A. Epple, C. Scanes and M. Stetson (Edi- tors, Progress in Comparative Endocrinology. Wiley-kiss, New York, NY, pp. 620-626.

Bromage, N., Randall, C., Duston, J., Thrush, M. and Jones, J., 1990~. The environmental control of spawning in salmonids. Proc. Satellite Symp. Applic. Comp. Endocrinol. to Fish Cul- ture, Almunecar, Granada, Spain, 22 May 1989. Publ. Univ, Barcelona, pp. 47-57.

Bulkley, R.V., 1967. Fecundity of steelhead trout (Salmo gairdneri) from the Alsea River, Or- egon, J. Fish. Res. Board Can., 24: 9 17-926.


Buss, K, and McCreary, R., 1960. A comparison of egg production of hatchery reared brook, brown and rainbow trout. Prog,. Fish. Cult., 22: 7-10.

Carillo, M., Bromage, N,, Zanuy, S,, Serranno, R. and Prat, F., 1989. The effect ofmodifications in photoperiod on spawning time, ovarian development and egg quality in the sea bass (ai- centrarchus hzbrax). Aquaculture, 8 1: 35 l-365.

Craik, J.C.A. and Harvey, S.M., 1984. Egg quality in rainbow trout. The relation between egg viability, selected aspects of egg composition, and time of stripping. Aquaculture, 40: 11 S-

134. Crim, L., Glebe, B. and Scott, A., 1986. The influence of LHRH analogue on oocyte develop-

ment and spawning in female Atlantic salmon, Sulmo salar. Aquaculture, 56: 13% 149. Cumaranatunga, R,, Bromage, N.R. and Springate, J.R.C., 1985. Atresia in the rainbow trout

(Salmo gairdner~). Proc. 7th Int. Conf. of the European Society for Physiology and Bio- chemistry, Barcelona, Aug. 85, A I, 20- 1.

Duston, J. and Bromage, N.R., 1986. Photoperiodic mechanisms and rhythms of reproduction in the female rEinbow trout. Fish Physiol. Biochem., 2: 35-5 1.

Duston, J. and Bromage, N.K, 1987. Constant photoperiod regimes and the entrainment of reproduction in the female rainbow trout (Salm~ gairdneri). Gen. Camp. Endocrinol., 65: 373-384.

Duston, J. and Bromage, N.. 1988. The entrainment and gating of the endogenous circannual rhythm ofreproduction in the female rainbow trout. J. Comp. Physiol. A, 164: 259-268,

Duston, J. and Bromage, N., 199 1. Circannual rhythms of gonadal maturation in female rainbow trout (Oncorhynchus mykiss) J. Biol. Rhythms, 6: 49-53.

Escafire, A.M. and Billard, R., 1979. Change in fertilizability of rainbow trout eggs Iefi in the abdominal cavity during the post ovulatory period. Bull. Fr. Pisc., 272: 56-70.

Follett, B.K., 1984. Birds. In: G. Lamming. (Editor) Marshall’s Physiology of Reproduction, 4th edn. Vol. 1. Churchill Livingstone, London, pp. 283-350.

Gall, G.A.E., 1975. Genetics of reproduction in domesticated rainbow trout. J. Anim. Sci., 40: 19-28.

Glebe, B.D., Appy, T.D. and Saunders, R.L., 1979. Variation in Atlantic salmon (Salmosular) reproductive traits and their implications for breeding programs. I.C.E,S. CM. 1979/M, 23, 11 PP*

Happe, A., Quillet, E. and Chevassus, B., 1988. Early life history of triploid rainbow trout. Aquaculture, 71: 107-I 18.

Hardy, R., 1983. Salmonid broodstock nutrition. In: R. Iwamoto and S. Sower (Editors), Sal- monid Reproduction. Washington Sea Grant Program, University of Washington, Seattle, WA, pp. 98-108.

Harris, L. and Griess, L., 1978. Trout nutrition and disease studies. Progress Report Dept. Na- tional Res. Colorado Division Wildlife F-28-R. 14,45 pp.

Henderson, N,, 1963. Extent of atresia in maturing ovaries of the Eastern brook trout Salvelinus jbntinalis. J. Fish. Res. Board Can., 20: 899-908.

Hiram, S., Yamada, J. and Kikuchi, R., 1955. Relation between chemical constituents of rainbow trout eggs and the hatching rate. Bull. Jpn. Sot. Sci. Fish., 2 1: 240-243.

Jones, J. and Bromage, N., 1987. The ihqfluence of ration size on the reproductive performance of female rainbow trout. In: D. Idler, L. Crim and J. Walsh (Editors ), Reproductive Physi- ology of Fish 1987. Marine Science Lab., St Johns, Newfoundland, p. 202.

Kate, T., 197% The relation between growth and reproductive characters of rainbow trout, Salrno guirdneri, Bull, Freshwater Fish Res. Lab., 25: 83-l 15.

Kate, T. and Kamler, Et 1983. Criteria for evaluz~ion of fish egg qyzzlity as. cxempli&d for Salmo gairdneri. Bull. Natl., Res, Inst. Aquacuh., 4: 61-78.

Kincaid, H.L., 1972. A preliminary report of the genetic aspects of I 50-day family weights in hatchery rainbow trout. Proc. S2nd Annu. Conf. West. Assoc. State Game Comm. Portland, OR, Oregon State Fish and Game @ommission, pp. 562-565.

BROODSTDCK MANAGEMENT AND EGG PRODUCTION 165

Kincaid, H.L., Bridges, W.R. and Vonlimbach, B., 1977. Three generations of selection for growth

rate in fall spawning rainbow trout. Trans. Am. Fish. Sot., 106: 62 l-428. Knox, D., Bromage, N., Cowey, C. and Springate, J., 1988. The effect of broodstock ration size

on the composition of rainbow trout eggs. Aquaculture, 69: 93- 104. Leitriz, E. and Lewis, R.C., 1976. Trout and salmon culture. Calif. Fish Game Fish Bull., 164,

pp. l-197. Lincoln, R. and Scott, A., 1983. Production of all-female triploid rainbow trout. Aquaculture,

30: 375-380. Morrison, J.K. and Smith, C.E., 1986. Altering the spawning cycle of rainbow trout by manip-

ulating water temperature. Prog. Fish Cult., 48: 52-54. Nakari, T., Soivio, A. and Pesonen, S., 1988. The ovarian development and spawning time of

Sui~~~~~u gairdneri reared in advanced and delayed annual photoperiod cycles at naturally flue- tuating water temperature in Finland. Ann. Zool. Fenn., 25: 335-340.

Nicholl:s., A.G., 1958. The egg yield from brown and rainbow trout in Tasmania. Aust. J. Mar. FresP-iwater Res., 9: 526-536.

Nomura, M., 1963. Studies on reproduction of rainbow trout (Sulmo gairdneri) with special reference to egg taking. IV. The fecundity or number and weight of eggs taken. Bull. Jpn. Sot. Sci. Fish., 29: 325-333.

Nomura, M., Sakai, K. and Takashima, F., 1974. The overripening of rainbow trout. I. Tem- poral morphological changes of eggs retained in the body cavity after ovulation. Bull. Jpn. Sot. Sci. Fish., 40: 977-984.

Orr, W., Call, J., Brooks, J., Holt, E. and Mainwaring, J., 1982. Effect of feeding rates on spawning performance of 2,year-old rainbow trout broodstock. Fish Dev. Centre Inform. Leaflet, US. Fish Wildlife Service, Bozeman, MT, No. 24,6 pp.

Phillips, A., Hammer, G., Edwards, J. and Hosking, H., 1964. Dry concentrates as complete trout foods for growth and egg production. Prog. Fish.-Cult., 26: 155-l 59.

Pitman, R.W., 1979. Effects of female age and size on growth and mortality in rainbow trout. Prog. Fish-Cult., 4 1: 202-204.

Randall, C. and Bromage, N., 1992. Short periods of continuous light provide a simple, cheap and predictable method for the production of out-of-season rainbow trout eggs. Aquaculture, 100: 172-173 (abstr.).

Randall, C., Duston, J. and Bromage, N., 1987. ?notoperiodic history and the entrainment of the annual cycle of reproduction in the female rainbow trout. In: D. Idler, L. Crim and J. Walsh (Editors), Reproductive Physiology of Fish 1987. Marine Science Lab., St. Johns, Newfoundland, p. 3 10.

Ridelman, J,, Hardy, R. and Brannon, E., 1984. The effect of short-term starvation on ovarian development and egg viability in rainbow trout. Aquaculture, 3’1: 133-l 40.

Rolcy, D., 1983. The effect of diet protein level, feeding level, and rearing water temperature on growth and reproductive performance of rainbow trout broodstock. Ph.D. Thesis, Univ. Washington, 289 pp. Diss, Abst. Int. V.44/04-B, 961.

Rounsefell, G.A., 1957. Fecundity of North American Salmonidae. US+ Dept. Interior, Fish. Bull. Fish Wildlife Service, 122, pp. 45 l-468.

Sakai, K,, Nomura, M., Takashima, F. and Oto, H., 1975. The over-ripening phenomenon of rainbow trout, II. Changes in the percentage of eyed eggs, hatching rate and incidence of abnormal alevins during the procss of over-ripening. Bull. Jpn. SOC. Sci. Fish., 4 I : 855-860.

Sandnes, K., Ulgenes, Y., Braekkrin, O.R. and Wtne, E, 1984. The effect of ascorbic acid SUP-

plementation in broodstock feed on reproduction of rainbow trout (Salmogai~dneri). Aqua- culture, 43: 167-l 77.

Satia, B., 1973. Growth and reproductive performance of rainbow trout (Su!mo gairdneri) to

different diets, Ph.D. Thesis, Univ. Washington, 98 pp. Satia, B.P., Donaldson, L.R., Smith, L.S. and Nightingale, J.N., 1974. Comparison of ovarian

166 N.BRGMAGEETAL.

fluid and eggs of University of Washington trout (Salmo gairdneri). J. Fish. Res. Board Can., 31: 1796-1799.

Scott, A. and Sumpter, J., 1983. The control of trout reproduction -basic and applied research OF hormones. In: J. Rankin, T. Pilcher and R. Duggan (Editors), Control Processes in Fish Physiology. Groom Helm, London, pp. 176-199.

Scott, D., 1962. Effect of food quantity on fecundity of rainbow trout. J. Fish. Res. Board Can., 19: 715-731.

Small, T., 1979. Trout eggs - look for size and service. Proc. 11 th Two Lakes Fish Symp. Oct. 1979. Romsey, England. Janusen Semites, Kent, pp. 127-l 32.

Smith, C.E., Osborne, M.D., Piper, R.G. and Dwyer, W.A., 1979. Effect of diet composition on performance of rainbow trout broodstock during a three year period. Prog. Fish-Cult., 41: 185-188,

Snedecor, G.W. and Cochran, W.G., 1980. Statistical Methods, 8th edn. Iowa State University Press, Ames, IA.

Springate, J.R.C. and Bromage, N.R., 1983. Egg production and broodstock management. Fish Farm., 6: 21.

Springate? 9-R. C. and Brompge, N.R., 1984a. Broodstock management: egg size and number - the “trade-or’. Fish Farm., 7 (4): 12- 14.

Springate, J.R.C. and Bromage, N.R., 1984b. Husbandry and the ripening of eggs. Fish Farm., 7: 22-23.

Springate, J.R.C. and Bromage, N.R., 1984~. Rainbow trout egg and fry losses: a check on quality. Fish Farm., 7: 24-25.

Springate, J.R.C. and Bromage, N.R., 1985. Effects of egg size on early growth and survival in rainbow trout (Sulmo gairdneri R. ). Aquaculture, 47: 163-l 72.

Springate, J.R,C., Bromage, NR., Elliot, J.A.K. and Hudson, D.L., 1984. The timing of ovulation and stripping and their effects on the rates of fertilization and survival to eyeing, hatching and swim-up in the rainbow trout (Salmo gairdtieri R. ). Aquaculture, 43: 3 13-322.

Springate, J.R.C., Bromage, N.R. and Cumaranatunga, R., 1985. The effects of different rations on fecundity and egg quality in the rainbow trout (Saltnogaivdneri). In: C.B. Cowey, A.M. Mackie and J.A. Bell (Editors), Nutrition and Feeding in Fish. Academic Press, London, pp. 371-391.

Takeuchi, T., Watanabe, ET., Qgino, C., Saito, M.,Nishimura, K. and Nose, T., 1981. Effects of low protein-high calorie diets and deletion of tract elements from a fish meal diet Irrr; reproduction of rainbow trout. Bull. Jpn. Sot. Sci. Fish., 47: 645-654,

Thorpe, JE:., Miles, M.S. and Keay, D.S., I984. Development rate, fecundity and egg size in Atlantic salmon (Sulmo salar L. ). Aquaculture: 43: 289-306.

Tyler, C., Sumpte r. J. and Bromage, N., 1987. Uptake of vitellogenin into cultured ovarian follicles of rainbow trout. In: D. Idler, L. Crim and J. Walsh (Editors) Reproductive Physi- ology of Fish 1987. Marine Science Lab., St. Johns, Newfoundland, p. 22 1.

Vladykov, V., 1956, Fecundity of the speckled trout (Suivelinusfontinalis) in Quebec lakes. J. Fish Res. Board Can., 13: 799-84 1.

Van Bayer, H., 1950. A method for measuring fish eggs. Prog. Fish-Cult., 2: 105-107. Watanabe, T., 198% Importance of the study of broodstock nutrition for further development

of aquaculture. In: c. Cowey, A. Mackie and J. Bell (Editors), Nutrition and Feeding of Fish. Academic Press, London, pp. 395-4 14.

Whitehead, 0. and Bromage, N., 1980. Eflects of constant long and short day photoperiods on the reproductive physiology and spawning of the rainbow trout. J. Endocrinol., 87: 6-7.

Whitehead, C., Bromage, N., Forster, J. and Matty, A., 1978. The effects of alterations in photoperiod on ovarian development and spawning time in the rainbow trout. Ann. Biol. Anim. Biochim. Biophys.. 18: 1035- 1053,

Broodstock management, fecundity, egg quality and the timing of egg production in the rainbow trout (Oncorhynchus mykiss

Documents