Utilization of Dietary Carbohydrates and Lipids by Salmonids Sexually Sterilized with 17alpha-Methyltestosterone

Two, twe lve week feeding tr ials were conduc ted to determin e the performan ce of j u venil e and yea rl ing rainbow trout CSalmo ga irdn er i) whic h wer e fed high-fat or high-carbohydrate diets. In each feeding trial and f or each diet group, 50% of the fish used had been previously tr eated with 17a-methly-t e stosterone, given orally at the onset of feedin g. These were called treated fish. Also , for each group, 50% were untreated rainbow trout (control s), kept and fed under id e ntical conditions ~~ treated fis h . All fish we re held in 100% flow-through systems an d fed tw ic e or three times daily. Fe0d requirement was determined as a percen t age of the total fish weight for each tank, accor di ng to water temperature .and average siz e of fish calculated a t each Th r ee replicates were used in the first feeding tri al and two in the second. Growth measurements Performance was determined through teed c0n version e fficiency, 9Ctual weight yain, actual length ~s in. relative weight gain and con dition f;;;.ctor Pla~ma glucose and ammonia levels, an d visc2ra l Two diets were For mulated in t he first trial , one high in lipid C24. 1%) with no carbohydrate add e d and the other

In the second trial three diets wer e for mul a t ed: o ne high in fat C24 . 1% lipi d), one high in sucrose (25.8%) and low in l ipi d (14.1%), and one high in molasses C42 . 0%) and low in lip id. All diets were iso-nitrogenous and iso-caloric.
The o b jec ti ve in these studies was to deter mine if a c or relation existed between performance of treate d or con tr o l rainbow t~out an d diet fed . diet performe d better than those fed the high-fat d iet.
groups with respec t to growth pa rameters, th ough not sig n iFicantly so in 1 ., a.i. i inst:::in c es;.
con ver ::= l on 1 n th E Arner i can eel ( Dt:~q an i and G.::i. Naga i an d Ik ed a (1973 ) sugg est t he i n feriority of carbohydrate as an ene rgy source in c arp an d al so demonstrated impa i red glucose u ti lization in these f i sh .
Hm·iever-, An cl er-so n et al (   Othe~ studies involving increasing proportion s of sucrcse and gelatinized maize starch in diets fo r rainbow t1'··uut lia.·v'E' !:Oh Dr.··!f"l t!""i,:.~t. t he1'··i:::: may i:JE· pui::,'.::.ibil.ities fo1'· CHU u5e i~ ~almonid diets. Pieper and Pfeffer (1980a) 1 in a stud y on the compar ati ve effi c ienc y of ut ili za tion uf gruss found that in r ainbow trout, sucrose was the most eff iclentl~ used energy source , with sunflow e r oil, gelatinize d sta rch and glucose Fo l lo wing in descending Fr om th is study it was concluded that glucose a n d  biologica l role of cortisol in fish and oth e r ver t ebra t es is to promote qluconeo ge nesis and enha n ce catabolic activity (Palmer , 1966;Stor er, 1967;Hendricks et 61. 19f3LJ.) .
By reducing blood pl as ma cortisol l ev els, insulin m2v increase , a ll owing fo r better utilization of bloo d glucose This ~or~ al so i ndicated th a t cor t isol was a n dffect ive gluconeogeni~ agent in the sh~rk.   1981;Macintosh et al., 1985 ) , ca r p ( Jensen et al., 1983; Rao et al. ,1 983> and s al mon id s (Do n aldso n and Hu n ter, 1982 ; Schreck and L i, 1983 ). Thi s tr eatment pro d uces s ex-inverted male (previously f emale s ) fis h wh ic h can then be mated to normal f males to p rod u ce a ll fe male pro g eny .
This has been achie~ed in At l a ntic sal mon ( Johnstone and Younqsnn . 1 19f34; . . This t echnique may be i mpl e men te d by fish farm er s to prod uce monosex cult ur es ne eded for spec if ic Sterility ha s been produced in salmonids b y feeding higher levels ~f 1 7aMT C40-60 mg/Kgl i n the di e t <Bill ar d ThesE? 3terile Fis h , after maturity, become easier to handle 3n d pr ove to be more hardy than untreated fish CS mith,1983).
lhe; 3lso ~how i mproved growth rates over un tr e a ted f i sh of Thes e author-s demonstrated e n hancement of thyroid activity in fish fed t7aMT-treated diets, as did Lone and Matty (1980)  17aMT was proven as a seven m0nth period (Fagerlund et al., 1979). I mp 1·· ovecnen t in growth and f eed con v ersion has also been demonstra t ed in rai nbow trout and Atlantic salmon with inc o rporation of 17aMT into diets CHiggs et al.

21
The use of 17aMT in fish culture operations is a hig hly fe asible one because this synthetic hormone i s i--·el.::-<t i -. ..-' e l y ine;-:pensive, on l y sma l l a mount~; are:· 1~e quired, they are used succe ssf ully in sex control studie s, and benefits with r espect to we ight g ain a nd feed conversion It has also been shown t ha t th is hormune is safe to use and is eliminated ( cleared) from tissues in a s:.hor·t pe1 '"iod o-f time af t1-::2 r· e::i.pplicatlon tias. CE•<:1sed ..  were held in a large tank with a 90% water reuse system at the East Farm Aquacult ure Center CEFAC> of the University of F:.i··1ode Isl=;_n d until needed f or-· e;.;ppr-imenta,l pu.1r·puses.
At the age of six month s for the first feeding trial, and ten months for the second feeding trial thes e sterilized f ish were then transferred, in appropriate number s, to smaller bioassay tanks which operated with a 100% water flow-through system . At this time, untreated (control) rainbow trout, which had been held in an identica l situation to the tre ated fish, were also transfered to the Steriliz ed C17aMT-treatedl an d control (non-treated) f i sh were of the same spawning group and therefore the same age at the time of the experiment. All experimental fish were held in a 100% water flow-through system throughout both feeding trial s . All fish wer e spawn ed , hatched and raised at EFAC and were f ed identi ca l die ts unti l the e xp eriments began. Treated fish we re distinguished from control fish by clippi n g of the a d ipose fin of treated fish prior to plac in g in experi menta l t anks for both feeding trials.
At the termina t ion of both feeding trials, fish were

Ex per i mental Design
After transferring fish from larger tanks to sm all er e x p er imental tanks and prior t o the start of bo th feed ing tri als , fish were a llowed to acclimate to new surroun di n gs for 10 to 14 days. Si x tanks C200 litre capacit y> were used for each feedin g trial. F or the first trial, 20 rainbo w trout CS. gairdneri) wer e placed in each tank, 10 of wh ich were tre ated f i sh a nd 10 control fis h . This was done to elim inat e any tank effect du ring the course of the trial .
Througho ut the first trial, which last ed 1 2 weeks , tanks 1-3 were fed a diet high in fat while tanks 4-6 were f ed a diet high in sucrose (as a digestible carbohydr ate> . For the second feeding trial, 24 rain b ow trout were placed in each tank, 12 of which were treated f i sh and 12 of which were control fish. The branding mechanism was easily built and the brands applied afte r fish were anesth esizeci Csee appendix !).
In the second feeding trial, a molasses Jie t was incorporated, with molasses providing the digestible CHO as opposed to sucrose. In this trial two tanks were f ed a high fat diet Clanks 1 and 4), two tanks were fed a high sucrose diet (ta nks ~ ~nd 5) and two tanks a high molasses During the sixth an d ninth week length and weight measurements, blood was taken f ro m two fis h (one treated, one control) µer tank and plasma obta in ed and analyze d for Results were again compared for each tank and each diet.
As with the first trial, up on termination (12 weeks) fish were weigh ed, measur ed. bled from the caudal vein with hep ari niz ed syringes, and sacrificed. Feed co nversjons, for the last growth period and overall, were calcula ted.
Anal ysi s wf growth data and calculation of overall fe ed conversions for the three feeds and tre ated ver sus con t rol Fi sh Nas done at nine weeks for the second trial. Tat and sucrose diets water was also added, and 31 alpha-cellulose was us ed as a f i ller in the molas ses d iet. Gr·a.nulated su gar· in the sucn::ise diet and "b l ack~,t.t·· ap " molas ses i n the molasses d i et were a dded i n an amount to provide an e qual number of c al o ries in both diets. All dietar y i ngredients (Table III) were mi x ed and e xtruded in to feeding pe llets of the approp ri ate size, as in trial #1.

Effects of Methyl-Testosterone Treatment
The effects o f MT treatment o n experimental f ish, as described in the methods and materials section, are shown in Al l control fish <100%> were fully deve l oped in beth tria ls and we re se x ed with ease.
ratio in control fish favored the female 1:1 .5, wh ile in treated fish which were identified the sex ratio was 3:1 .
Tota l effects of treatment (Table V> are

41
:t -relative gain = gain <grams> / initial weight <grams) 2 = protein conversion = I. protein efficiency *fish fed sucrose diet had significantly better CP<0.05) actual weight gain, length gain and feed co nversion than did those fed high-fat diet +treated fish on the sucrose diet had significantly improved <P<0.05> actual weight gain over treated and control fish fed the high-fat diet; sucrose fed treated fish had significantly impro ved CP<0.05) relative weight gain over all other groups ++ ~ treated fish fed the high -fat diet had the poorest feed conversion; the remaining groups were not significantly different CP<0 .05)

Figure 2
Actual gain Cgrams) of treated and co n trol ra i n bow trout fed high-fat and sucrose diets for 12 week s .

I-50
:c Protein conversion efficiencies were naturally also better in treated fish on the sugar diet because all of the diets were iso -nitrogenous (e qu a l protein levels).
Re lat iv e growth or gain <gai n/initia l weight>, shown in ,.

ACTUAL GAIN
52 grew better (taking into account initial size) than did treated fish on the high fat diet. Treated /su crose fish also gained better than control fish on both sugar and high fat diets. Figure 4 sho ws r elative growth over time for each group. Figure 5 s h ow s graphically the interaction effects of some of the growth parameters. Condition factor was calcu lat ed for each group based on mean data and was not found to be sig nific antl y different between diets or treatments.
Individu al measurements were used in determining visceral fat levels (% body weight) and blood plasma glucose differences between groups. Results indicated that the blood glu cose of treated fish at termination was significantly l ower <P<0.05) than that of control fish in bo th the hig h fat and sucros e diet groups. body weight> are tabulated in Table VII and illustrated g raphi cally in Figure 6.
No correlation wa s found between blo od plasma a mmonia levels and blood pla s ma glucose le vels or between pl asma ammonia levels and treatme nt or diets.
There was no significant difference between fish blood ammonia on different die t s, trea t ment or any interaction therf.O>of ..

Trial II-Sucr-ose ~ Molasses ~ High Fat Diets
Actual weight gain and length gain were determined for each fish and mean values were calculated along with feed conversion for both treated and contr-ol fish on each diet.
.   Table VIII. Figure 7 illustrates weight gain over time for control and treated fish within each diet. Feed con v ersion data is represented in Figure 8.
Mean relative gain (wei~ht gain/initial weight) for treated and control groups fed the three diets was calculated in the second trial by averaging individual relative gains for each fish <Table VIII>. After nine weeks of feeding, there was a signif ic ant difference CP<0.05) in relative growth between both feed s and treated versus control groups. Interaction of the two variables (feed and treatment) was also found to have a significant effect on relative gain at P<0. 01

1.183++
C:::.7.6%) 1 = f inal weight at 9 weeks 2 -relative gain = gain (grams)/initial weight (grams) 3 = protein con ver sion = % protein e f ficiency *sucrose and molasses diets gained significantly better <P<0.05) than did the high-fat diet ** = all diets were significa n tly different CP<0.05) from the other two diets in length gain a nd relative weight gain (molasses and sucrose were s uper ior) += treated fish fed mo l asses showed significantly improved relative gain over all other g r oups; sucrose > high-fat ++== molci.sses and sucr-ose diets g«:i\/e signi·ficantly bE~ttE!t·· CP>0.05 ) fe e d conversions than the high-fat diet

values.
Blood plasma glucose values of treated fish were not significantly different from those of control fish.
Measurement of visceral fat as a percentage of body weight resulted in significant differences <P<0.05) between feeds with only the sucrose diet being significantly lower visceral fat than the other two diets. There were no . ., in ,.
significant differences in visceral fat levels between treated and control groups, but treated fish did have lower mean levels of visceral fat (1.021%) than did non-treated f i sh ( l . 0286 ~~) .
Average results for visceral fat % and blood plasma glucose levels are shown in Table IX and in Figure 11. Ammonia levels in blood plasma showed no correlation with any other factors, with no significant differences in levels between treatment or diets .
In a related feeding study, using three c ommercial pelleted diets which contained varied level s of carbohydrate~ treated and control rainbo w trout were individually marked usi n g a f r eeze-branding techniq u e and fed for 12 weeks under stringent controls <Smit h and Ahern, 1985). Resu l ts from the study, agreeing wi t h some of those obta i ned from this thesis researc h , showed that treated fish outperformed controls in all parameters measured and did the best on the diets highest in carbohydrates.
Table IX -Plasma glucose and visceral fat levels of trea ted and control rainb ow trout after feeding for 9 weeks with high-fat, sucrose and molasses diets -Trial #2 Hi  Relative gain (ga in /in itial wt) of treated an d control rainbow trout fed high-fat, sucrose and molasses diets fo r 9 weeks. Trial #2.       (Jalabert et al., 1975;Johnstone et al., 1978;Harbin et al., 1980;Donaldson and Hunter, 1982;Schreck and Li, 1983). Control fish, on the other hand, were 100% identifiable as mal es or females and had a more balanced sex ratio Cl:l.5) as compared to treated fish, and favored the fe mal e. This may indicate sex reversal from female to male when complete sterility did not occur with MT treatment. Also in agreement with previous research <Higgs et Rao et al., 1983;Smith, 1983), i t a ppeared that overall, tr eated fish performed better in b ot h feeding studie s as compared t o 75 controls, long af t er eli mina ti on of MT fr om the diet. In a more detail e d study on the long term effects o f MT treatment in fish, it is sug ges ted that other me a su r emen ts might be made, such as ci rc ulatin g seru m testosterone a nd plasma gonadotr op in CGthl le v els and cal culation of gonadosomatic index differences between tr eate d and cont rol groups <Billard et al., 1981;Sundararaj et al., 1982> . These perame te rs of sexual reprod uc tive potential would give a good indication of how well the MT t rea tment worked.
In a s tud y re la ted to th e feeding trials of the present study using three comercia l pe ll eted diets fed to treated and con trol fish, results i nd icate d the superiority of treated fish over co n trols fo r all growth me as urements.
Th e use of glucose , s ucr ose a nd more c o mplex CHOs in sal monid diets has given mi xe d r esults .
Investigators have d emonstrated that more complex CHOs such as starch are poorl y digested by salmonids <Edwar ds e t Spannhof and Plantikow, 1983), resulting in poor growth and feed conversion.
Other res earche rs though, have shown that restriction of intake and gelatinization of starch allows for better digestion and hen ce be tter results <Pi e per and Pfeffer, 1980 b;Berget an d Breque, 1983 In th e first feedin g tr i al (sucrose v s high-fat), though equal amounts of metabolizable ener g y were provided in each diet, the high-sucro se diet outperformed the high-fa t diet in all respects. Fish <treate d and control combined) f ed the high-sucrose diet s ho wed significantly improved actual weight, relative weig ht gain ( ga in/initial weight>, length gain and feed conv ersions <Tab le VI; Fi gure s 2 ,3,4).
In sucrose fed fish, both a ctual we ight gain and relative weight g a i n we re significan tly enhan ced .
Relative In the second feeding tri al~ where molasses was incorpor at ed as a n alternate digestible CHO source, both of the high CHO diets demonstr ated significantly increased act ua l weight gain and actual length gain than the high fat diet after nine weeks of feeding rai nbow trout (Table VIII The mol a sses diet al so showe d higher rel ativ e gai n than t h e su c rose diet in this regard , due to the superior growth of treated fish on the molasses diet. Therefore with respect to relati ve gain, molasses Cl.53 Sucrose has been demonstrated to be a superior feed additive over glucose and starch also in the tilapia, Dreochromis niloticus <Anderson et a~. ,1984). These results along with those given in other studies using rainbow trout (Pieper and Pfeffer, 1980a)  containing only 50-60% sucrose, must be added in higher amounts to fiah diets tu provide equal caloric value. But this waste product of the sugar refining industry is relatively inexpensive in most areas and is palatable to the fish, making i t also practical for aquacu ltural use.
The use of the s e quality CHO ingredients along with the cheaper, high quality cooked ur uncooked protein mixes which Robinson et al (1985) investigated, might provide the aquaculture industry with an inexpensive and high performance fish feed.
The availability and cost practicality of sucrose, molasses and other highly digestible CHOs such as brewers so lids, makes the possibilities for use of these products as fish feed ingredients near limitless. The evidence submitted here and in other studies has shown the advantages of using suc h additives in any fish diet formulation.
Diets using these ingredients are easily prepared, can produce a good feeding pellet, and will provide physi.cal properties su perior even to high -fat 83 p reparations . Incorporation of these products produ ces a feed which i s highly palatable to the fish and gives more than ade qu ate results as compared to conv enti o nal pelleted feeds used today.
The effect of MT treatment on f is h used in these feeding trials appeared to elicit the desired response to the differen t diets.
In a l l cases growth parameters indicated that treated fish di d much better on a high-CHO diet than on a high fat diet.
In the first feeding trial, MT -treated f ish had much poorer mean feed conversions when fed the high-f at diet <1.50> than when fed the sucrose diet Cl.24 ) and also did better than control fish fed either the high-fat or sucrose diets <1.28).
The same was true with respect to actual weight gain in the first trial. Treated fish fed the sucrose diet showed significant ly better gains than any other group. Plasma glucose levels were significantly lower in treated fish on both diets, which suggests improved glucose uptake at the cellular level.
Percentage of v i scera l fat was significantly lower for treated trout (~1.20%) than for control fish (~1.68%), another indication that the treatment given earlier in the f ish 's life has had so me effect on dietary CHO util ization, and the laydown of visceral fa t. The first feeding trial clearly indicated an advantage of using MT-treated trout over non-treated trout when feeding a digestible carbohydrate diet.

84
In the second feeding trial, results did not show such a clear-cut advantage of treated fish over untreated but, generally growth parameters were bett e r for MT-treated fish than controls.

CHO.
This was especially evident on the diet highest in Treated fish also had lower levels of visceral fat (as % of body weig ht) than control fish fed the same diets.
Explanations of why treated (sexua ll y-sterile> fish perform better on a high-CHO diet ha ve been alluded to in the introduction and literature review of this thesis.
Treatment with MT at the onset of feeding in young fish may have affected the normal endocrinological controls in these fish. Along with possible blockage of normal testosterone production in these fish, pre ven ting sexual maturity, it is possible that normal cortisol production is somehow  CHO. This stud y also shows that a high-sucrose or high-molasses diet will produce an enhanced growth response than a high fat diet of comparable energy value. It also dispells the belief that salmonids require, and only do well on a high-protein diet to the exclusion of all else. Lower leve ls of crude protein which contain adequate amounts of essential amino acids may certainly be used in conjunction with higher levels of digestible carbohydrates. Also~ the use of freeze branding in the second trial of this study has shown this method of mar~ing rainbow trout to be a valuable tool in obtaining individual statistics f or analysis of experimental data (see Apendix I>. Feeding studies and other research using fish, which ex tend over a 12-20 week period, may be analyzed individually to acco unt for differences due to diet and/or treatment.
With proper technique, thi s method provides clear brands for at least 16 weeks and would preclude infection and disease problems encountered in con ven tional tagging methods.
In future such MT--treated versus control type studi es, it is sug gest ed that ex perimental design takes into account several aspects which have been overlooked in this study.
Perhaps, because treated fish have been shown to be more docile and less aggressive than controls of the same age and spawning group (S mith? 1983), as part of the study, treated fish st1ould be maintained apart from c ontrols in order to determine t1ow this aggression aspect would affect growth performance.
It is also recommended that, as has been done in other carbohydrate feeding studies, liver weight and hepatic lipid and glycogen le v els should be measured, in order to determine the effect of diet on this organ with respect to CHO utilization. In order to determine positi vely blood parameters which control CHO utilization in treated or control fish? perhaps plasma cortisol, testo sterone, thyroxin and/or gonadotropin <GTH> l ev els should be measured. Muscle tissue or whole body a nalysis of fish a fter termination of each trial could be done to determine the effect of each diet on quality and quantity of body flesh. Dressed ,carcass weight could also be determined at termination as a measurement of feed and protein conversion. One may conclude that a l though good evidence of the superiority of diets hig h in digestible carbohydrate, and better growth performance of treated fish has been demonstrated in this study, further research into the nutritional and hormonal aspects of fish culture must be continued.
To help meet the increasing demands for high -protein foods in the modern world at minima l cost, these nutritional and hormonal applications may be instituted in fish culture operations.
Appendix I -The use of a freeze-branding mechanism as a means of marking rainbow trout for later individual identification.
In a preliminary feeding trial which was not included in this thesis report, brook trout (Salvelinus fontinalis) were individually marked with plastic, numbered tags at the start of the trial.
Tags were manually imbedded just ventral to the dorsal fin into the dorsal musculature using a tagging gun.
The needle of this gun was injected under the skin and into the muscle and anchored there thr oug h a twisting action of the gun. Within three weeks of the taging high mortality began to occur. The area around the tags became infected with bacteria CAeromonas spp.> and ulceration and necrosis of the tissue resulted.
Tags often fel l out , and by the en d of the feeding trial at least 50% mortality occured in each tan k.
Primary ba cteria l infection was often fallowed b y secondary bacterial or fungal <Saprolegnia spp.) A less stressful, safer and more effective method of marking must be used to insure more dependable resu lts with no mortality problems.
In the second feeding trial of the present thesis research, a freeze-branding procedure using liquid nitrogen was utilized to individually mark all of the fish in each tank. The f reeze branding apparatus was easily constructed and inexpensive to make. To the end was solde red a solid copper "T" i•Jh i ch trJas L'.sed as the brand. This "T" was 0.8 cm across the top and 0.7 cm from top to bottom.
During the branding operation, the section of the rod inside the reservoir c~10.0cm> was completely immersed in liquid nitrogen. The holding capacity of the brass reservoir <>1.0 litre) allowed for effective branding of more than 140 fish. The whole apparatus was mounted on a plywood platform making it portable and safe ta use.
At the start of the second feeding trial used for this thesis, r2inbow trout were individual ly branded fo~ later ide ntification using the following procedure.  Figure 12 shows a recognizable brand on a rainbow trout 6 we eks aft e r the brand was applied.