Short-term profiles of plasma gonadotropin and estradiol-17 beta levels in the female rainbow trout, from early ovarian recrudescence and throughout vitellogenesis.

At various stages throughout the annual reproductive cycle, female rainbow trout (Salmo gairdneri) were fitted with a catheter in the dorsal aorta. They were bled via the catheter at frequencies of once every 30, 60, or 240 min over periods of 5 to 24 hr. Gonadotropin (GtH), estradiol-17 beta (E(2)17 beta). and estrone levels were measured in the plasma samples. At early ovarian recrudescence (March), short-term (1-2 hr), high-amplitude (delta GtH = up to 100 ng/ml), episodic pulses of GtH were recorded in the plasma of 12 of the 26 studied fish. In the others, GtH levels remained low and constant. No synchronization was found among the individual GtH profiles. E(2)17 beta levels in the same fish were low and constant while estrone was not detectable. At early stages of exogenous vitellogenesis (June), plasma GtH (1-3 ng/ml) and E(2)17 beta (0.5-1.5 ng/ml) levels were low and constant. At advanced stages of exogenous vitellogenesis (September-October), fluctuating GtH levels were found again in most of the females; basal GtH concentrations were only slightly higher than those recorded in June. The fluctuations consisted of short-term (1-3 hr) random GtH pulses of moderate amplitude (delta GtH = up to 5 ng/ml), occurring at a relatively high frequency (up to 5 per 12 hr). Although no regular synchronous daily pattern of GtH was noted, most of the GtH pulses were observed during the photophase and early scotophase. The appearance of GtH pulsatility during exogenous vitellogenesis was accompanied by a large increase in plasma E(2)17 beta up to levels ranging from 6 to 30 ng/ml. In contrast to the GtH profiles, the E(2)17 beta profiles showed continuous and progressive variations, superimposed the abrupt GtH pulses, and a high degree of synchronization. E(2)17 beta levels increased during the photophase and reached maxima toward and during early scotophase.

The seasonal variations in plasma gonadotropin (GtH) and sex-steroid levels were thoroughly studied in the female trout (Salmo gairdrzeri). A careful examination of the various studies reveals some discrepancies amongst them. A transient increase in GtH levels has been observed by some authors at early stages of ovarian development (spring) in the rainbow (Breton et al., 1985) and brown  trout, whereas others have reported constant or only slightly varying levels of the hormone (rainbow trout: Scott and Sumpter, 1983;Sumpter et al., 1984). At the same time, estradiol-17B (E,17B) levels have been reported as low and constant (rainbow trout: Lambert et al., 1978;Whitehead et al., 1978a, b;Scott et al., 1980;Van Bohemen and Lambert, 1981) or increasing steadily (brown trout: rainbow trout: Sumpter et al., 1984). During vitellogenesis plasma GtH levels have been reported as either (a) increasing moderately (brown trout: Crim and Idler, 1978), (b) remaining constant (rainbow trout: Scott and Sumpter, 1983), or (c) initially increasing and then declining continuously (rainbow trout: Bromage et al., 1982a, b;Whitehead et al., 1983). Plasma levels of E,17P and testosterone rise considerably during vitellogenesis and decline progressively from its final stages (brown trout: rainbow trout: Fostier et aZ., 197X;Scott and Sumpter, 1983;Scott et a/., 1980;Bromage et al., 1982a, b).
One possible explanation for the abovementioned discrepancies might be the fact that the long term (seasonal) profiles only partially reflect the hormonal function. In fact, in the majority of the mammalian species studied in this respect, the fundamental signals composing the hormonal message are generally of very short duration, i.e., from a few minutes up to a few hours (see review in Knobil, 1980;Lincoln and Short, 1980;Brinkley, 1981;Desjardins, 1981). Similarly, daily variations in the circulating levels of the reproductive hormones have been reported in different fish species. Daily cycles in plasma GtH levels have been demonstrated in the female goldfish, Carassius auratus (Breton et al., 1972;Peter, 1978, 1980a;Vodicnik et al., 1978; and in the male and female of the common carp, Cyprinus carpio; the silver carp, Hypophthalmichthys molitrix; and the grass carp, Ctenopharyngodon idellus (Pan et al., 1980). Circulating levels of testosterone, E,17@, and estrone have been shown to fluctuate during the day in the catfish, Heteropneustes fossilis (Lamba et al., 1983). In salmonids, data on circadian changes in the levels of reproductive hormones are rare. Our previous reports (Zohar, 1980;Zohar et cd., 1982a) have presented some data showing that in the female rainbow trout, daily (short-term) fluctuations in plasma GtH levels do occur and are superimposed on the seasonal variations of the hormone.
The aim of the present work was to study, at some characteristic stages of the reproductive cycle of the female rainbow trout, the short-term (circadian and ultra-dian) profiles of plasma GtH and main sexsteroid levels. Since our previous data (Zohar, 1980;Zohar et al., 1982a) bad indicated a lack of synchronization between fish in the GtH secretion pattern, we studied hormonal profiles of individual free-swimming trout, fitted with a catheter in the dorsal aorta. The present paper describes short-term profiles of GtH and E,17P from early ovarian recrudescence and throughout vitellogenesis, whereas our following one (Zohar et al., 1986) describes short-term profiles of GtH and 17a-hydroxy, 20P-dihydroprogesterone at the periovulatory period.

MATERIALS AND METHCI (a) Animals and Stocking Conditions
Female trout, aged 3 to 4 years and weighing I .5 to 3 kg, were used in the present study. All of them had reached sexual maturity at least once before. About 2 to 3 months before the beginning of the acclimation to experimental conditions, fish were transferred from farm to indoor tanks supplied with recycled water, Then, 3 to 4 weeks before the initiation of the experiments the fish were transferred into tanks (70-150 liters) in which they were kept individually. Fish were exposed to a natural photoperiod. Water temperature was maintained at 10 2 2" between December and May and at 15 i 2" between June and November. Fish were fed with artificial food granules at a daiiy ratio of 1% of their body weight. Feeding time was between 1000 and 1200 hr.
(6) Dorsal Aorta Catheterization At various times of each year over a 3-year period, catheters were implanted in the dorsal aortas of lo-20 female trout. Only fish which had showed active feeding behavior for at least 1 week were catheterized. The procedure for implanting a catheter in the dorsal aorta has been described in detail by Zohar (1980). The same study (Zohar, 1980) and Bry and Zohar (1980) showed that the resumption of normal feeding behavior after catheterization reflected the recovery of fish from the stress situation related to the surgery. In such fish, the surgery and the presence of the catheter in the dorsal aorta did not affect gonadotropin levels in the blood or ovarian function (Zohar, 1980).

(c) Blood Sampling
Before catheterization, 0.3-0.5 ml of blood was removed from the caudal vascuiature of every fish. Hor-mone levels in these samples were compared to those recorded in blood sampled successively via the dorsal aortic catheter during experimentation, as described by Zohar (1980). After catheterization, fish were allowed to recover for at least 3 days before further bleeding.
In agreement with our previous findings (Zohar, 1980;Bry and Zohar, 1980), only fish which resumed normal feeding behavior after surgery were selected for experimentation. Individual free-swimming catheterized trout were bled repeatedly via the catheter at frequencies of once ever %, 1, or 4 hr over periods of 5 to 24 hr. In order to bleed the fish, the cannula, its stopper (a sealed syringe needle), and its cork floater were gently removed from the water. The cannula was unplugged and fitted with a l-ml syringe. A volume of 250 (~1, corresponding to 50 (~1 more than the dead volume of the catheter, was removed and discarded. The cannula tip was then fitted with a clean l-ml syringe previously rinsed with lithium heparinate solution (loo-150 iu of heparine in 0.7% of saline). A volume of 300 to 400 pl of blood was withdrawn and transferred to ice-cold 3-ml tubes. At the end of each sampling, the blood tilling the catheter was pushed back by 200 pl of lithium-heparinate saline solution (as above) and the cannula was plugged. Bleeding during the dark period was made using a dim red light to which wavelength salmonids adapted to dark are not sensitive (Hanyu and Tamura, 1978). Blood samples were centrifuged (15 min at 5000 rpm and at 4"), and the plasma was stored at -30" until further analysis.

(d) Determination of Ovarian Developmental Stages
At the end of each experiment, the females were sacrificed and their ovaries submitted to histological analysis. We distinguished three major ovarian developmental stages (based on Van den Hurk and Peute, 1979): I. Early ovarian recrudescence (March). Ovaries contained isolated or grouped oogonia, oogonia undergoing mitotic divisions, oocytes at different stages of the prophase of the first meiotic division, and different proportions of oocytes undergoinng previtellogenie growth, endogenous vitellogenesis, and the very beginnings of exogenous vitellogenesis. Previtellogenie oocytes (50 and 500 pm in diameter) were characterized by the presence of Balbiany bodies. Their nuclei were relatively big and many nucleoli were scattered at its periphery. Their follicular layers were not completely organized. Oocytes which undergo endogenous vitellogenesis (400-900 pm in diameter) were characterized by the successive appearance of two structures: the vitellin vesicles (the future cortical alveoli) which appeared first in the ooplasm periphery and spread later toward its center, and the lipid globules which appeared in the perinuclear area, fused, and gained size. The zona radiata and the follicular layers were already organized in such oocytes. Oocytes undergoing early exogenous vitellogenesis are described below.
2. Early exogenous vitellogenesis (.&tie). Ovaries contained mainly oocytes undergoing initial stages of exogenous vitellogenesis. Such oocytes (which measured 1 to 2 mm in diameter) were characterized by the progressive appearance, in the periphery of the ooplasm, of vitellin granules.
Ovaries contained mainly vitellogenie oocytes with diameters of up to 4-5 mm. They were characterized by the massive appearance of vitellin granules all over the ooplasm. These granules coalesce to form spheric structures (vitellin spheres) which occupy most of the oocyte volume.
(e) Hormone Measurements Gonadotropin was measured by radioimmunoassay (RIA) according to Breton and Billard (1977). The antibody used was anti-trout GtH at a final dilution of 1:130,000 to 1:200,000. A highly purified salmon GtH  was used for the standard curve and for the radioactive labeling. Each unknown sample was measured in triplicate. All samples from a single experiment were measured in the same RIA run. GtH concentration in most of the samples was measured twice, in two different RIA runs, in order to confirm the existence of fluctuations in the hormonal concentration and to eliminate any error due to the measuring technique (see also Zohar, 1980). The sensitivity of the RIA varied between 15 and 30 pg GtH per tube which corresponds to 0.3 and 0.6 ng GtHiml of plasma. The intraassay variability (Table 1) was estimated by repetitive measurements, in each of the RIA runs, of the GtH concentrations in different pools of plasma. These pools were sampled from fish at different physiological stages and contained various GtH levels, covering the entire range of the standard curve.
The radioimmunoassays of estradiol-17S and estrone were carried out according to Fostier et al. (1978), except that bound steroid was precipitated with polyethylene glycol . Both steroids were measured after being extracted from the plasma and separated from each other on an LH-20 column. Unknown samples were measured in triplicate. The estradiol-17B antibody was used at a final dilution of 1:50,000, and its degree of cross-reactivity with other steroids was 11% for estrone, 9% for 16-ketoestradiol-17B, 8% for 16-epiestriol, and less than 0.5% for estriol, estradiol-17a, testosterone, 1 l-ketotestosterone, and 17a-hydroxy,20@dihydroprogesterone. The estrone antibody was used at a final dilution of 1:4800, and its degree of cross-reactivity was less than 0.8% for estradiol-17B, estriol, testosterone, II-ketotestosterone, and 17n-hydroxy,20l3-dihydroprogesterone. The sensitivity of both radioimmunoassays was approximately 10 pgitube, and the intraassay variability is given in Table 1.

(f) Analytical Methods
The mean, coefficient of variation (CV) and variance were calculated for all the values composing each of the individual hormonal profiles. A second (hypothetical) variance was calculated for each profile, based on the highest CV obtained for a plasma pool (Table 1) having a hormone concentration close to the mean of the values composing the tested profile. This hypothetical variance (and CV) is thus the expected one, supposing that the observed hormonal fluctuations reflected only the variations of the RIA. The real and hypothetical variances were compared by x2 test. On the basis of this comparison, all profiles having a real CV at least twice as large as the hypothetical one were considered as fluctuating significantly, due to physiological reasons. These fluctuating profiles showed a highly significant (P < 0.00s) larger real CV than a hypothetical one. For the nonfluctuating profiles (real CV S 2 x hypothetical CV), the mean of all values composing each of them was considered as the "basal level." We distinguished two types of fluctuating profiles: (1) Irregular profiles, showing random, episodic fluctuations of hormonal levels. In this group there was no synchronization between individual profiles of animals representing the same physiological stage; (2) Regular profiles, showing gradual and continuous changes in the hormonal levels. In this group there was a certain synchronization between profiles of individual animals. The two types of profiles were further analyzed by different methods.
(2) Regular profiles. A global analysis of the profiles of all the animals representing a given ovarian developmental stage was realized by a two-factorial analysis of variance (the two factors analyzed were number of animals and sampling time). When significant differences were found, the analysis was continued by the Duncan test (Bliss, 1967). This enabled the identification of significant differences, during the experimental period, between groups of mean hormonal levels.

RESULTS
(a) Early Ovarian Recrudescence: (Figs. 1 and 2) In trout which were sampled every 4 hr for 24 hr, we observed in five of the nine fish studied significant short-term increases in plasma GtH levels (Fig. 1). Hn two of the femaies (Nos. 113 and 115), the amplitude of the increases reached around 100 and 50 rig/ml, respectively, whereas in the others the amplitude varied between 4 and 8 rig/ml. No synchronization was observed among the individual profiles. In . Females were bled every 4 hr over a period of 24 hr. Individual GtH profiles with significant fluctuations in hormone levels, and the mean of profiles (n = 4 fish) with stable GtH levels, are presented. The number of each female is indicated. The statistical significance of the hormonal fluctuations is indicated by the following symbols: for GtH, fi = nonsignificant and Ic = significant; for E,17B, W = nonsignificant and T = significant. The symbol * situated above a peak indicates that it is significantly higher than the basal level.
four females, GtH levels remained constant. Four of the GtH peaks were observed in the middle or toward the end of the photophase, whereas another peak occurred during the scotophase. In the fish in which they were measured, estradiol-17l3 levels were found to be low and constant, whereas estrone levels were below the detection limits of the RIA (0.15 rig/ml).
Bleeding fish every hour for 12 hr, during the day (for one group) or throughout the night (for another), showed short-term increases (pulses) in plasma GtH levels in seven of the 17 sampled females (Fig. 2). In the other fish, GtH levels remained con-stant. The amplitude of the GtH increases was relatively large in some of the cases, and ranged from 5 to about 100 rig/ml. In most of the females, a GtH pulse included only one point (sample) after which GtH returned to its basal level within 1 hr. In one female (No. 122), in which the GtH peak value was the highest (118 rig/ml), GtH returned to its basal level more gradually, through an intermediate value observed 1 h after the peak.
Although no synchronization was evident among individual profiles, six of eight significant GtH pulses occurred during the night, whereas the others occurred in the late photophase (Fig. 2). In all the fish ex-(6) Early Exogenous Vitellogenesis: June hibiting GtH pulses, only one pulse was (Fig. 3) detected throughout the 12-hr sampling period, with the exception of one female, for At this stage of ovarian development, which we observed two significant GtH GtH levels were low (l-3 rig/ml) and relaelevations. In three females showing GtH tively stable. When females were sampled pulses, we measured low and constant once every 4 hr for 24 br (Fig. 3a), we ob-E,178 levels, whereas es&one was not de-served slight fluctuations in GtH levels in tectable (Fig. 2, female Nos. 122-124).
two of six fish, whereas in the others these a. levels remained constant. In all females sampled every 1 hr for 10 hr (Fig. 3b), GtH levels were low and stable. When measured, E,17P levels were found to be low and constant (Fig. 3b). Due to space limitations, we include here only a representative number of individual hormonal profiles out of nearly 50 made on fish at this stage of ovarian development.
Of the 47 analyzed individual GtH profiles of females undergoing advanced exogenous vitellogenesis, 29 exhibited significant daily fluctuations.
In Fig. 4 are shown individual GtH profiles of nine fish which were bled hourly for 9 hr. For two females (Nos. 145 and 147), two GtH profiles are shown which were determined in two separate RIA runs. Seven of the nine GtH profiles fluctuated significantly. In all these cases, one to two significant GtH pulses were found during the 9-hr sampling period.
In Fig. 5 are shown individual profiles of GtH, E,17P, and estrone in eight females which were bled once every hour for 12 hr. Six of the GtH profiles fluctuated significantly, whereas in another one (No. 155) fluctuations were very close to being significant. In four of the females, significant No synchronization existed among the individual GtH profiles of fish bled hourly. However, we noted 13 significant GtH pulses in the eight females which were sampled between 0800 and 2000 hr (Fig, 5), and only six significant pulses in seven females sampled between 2000 and 0800 hr (individual curves not shown). The lack of synchronization between the individual GtH profiles is reflected by their mean profile (Fig. 6, low curve), which masks the existence of the GtH pulsatility and shows very constant GtH levels over the 24-hr sampling period. Figure 7 shows GtH profiles in eight females sampled once every 30 min for 5% hours. In this case, significant fluctuations in GtH levels occurred in five of the females. Due to the more frequent samplings, most of the elevations in GtH levels included at least two values.
The analysis of all GtH profiles which we recorded showed that when oocytes undergo advanced exogenous vitellogenesis, the dynamics of GtH secretion differs from that found at earlier stages of gonadal development. One to five GtH pulses occur over a period of 12 hr; their duration ranges from I to 3 hr and their amplitude reaches around 4 rig/ml.
As far as estradiol-176 is concerned, its mean level was much more elevated than at earlier stages of ovarian development (Fig.   6, Student t test, P < 0.01). The levels of this steroid fluctuated significantly luring the day in eight of 11 studied females (Fig,  5 for the profiles of seven females sampled between 0800 and 2000 hr). The pattern of the individual E,I7@ profiles was different from that established for the GM. E,l7@ levels did not change abruptIy, but revealed continuous, relatively synchronized daily variations overriding the GtH pulses (Fig.  5). In the six fish showing fluctuating profiles out of the seven sampled from 0800 to 2000 hr, E,17P levels were lower in the first part of the photophase, increased progressively later on, and reached maxima during the later part of the photophase or at the beginning of the scotophase. In some of these females, E,17/3 tended to decrease after reaching maxima. This tendency continued in fish sampled hourly between 2000 and 0800 hr (curves not shown). The relative synchronization between the individual E,17l3 profiles is reflected in their mean profile (Fig. 6, upper curve), which shows a progressive, highly significant increase of E,17l3 during the day, from min-imal levels at 0800 to 1000 hr up to maxima reached between 1500 and 1700 hr, followed by an initiation of a decline. Plasma E,17P profiles were recorded only in four females sampled between 2000 and 0800 hr. These profiles (not shown) exhibited relatively high individual variation, and the calculation of their average was thus meaningless. More fish should be studied in order to confirm the E,17P pattern during the nocturnal part of the 24 hr.
We measured estrone levels in five females (Fig. 5). Those levels were very low, close to the detection limit of the RIA. In some of the females (Nos. 156 and 160), the estrone profiles were parallel to those of E,17P. 12 14 16 18 20 22 24 2 4 6 8 Time of day (hour) FIG. 6. Mean profiles (x 2 tSE) of plasma levels of GtH (lower curves-24 hr) and of E,17@ (upper curve-12 hr) in female trout at advanced exogenous vitellogenesis (October). The mean profiles were established from individual profiles of fish sampled every hour over a period of 12 hr (Fig. 5 for the group sampled from 0900 to 2100 hr). **Significantly different at P < 0.01.

DISCUSSION
Annual changes in GtH levels in the female rainbow trout have been described previously (see introduction).
The present study together with our following one (Zohar et al., 1986) reveal that in addition to the seasonal variations in the hormone concentration, important modifications occur in the pattern of the short-term profiles of GtH levels throughout oocyte development .
A transient GtH elevation has previously been observed in the circulation of the female rainbow and brown trout undergoing ovarian recrudescence. This elevation occurred either in spring (March) in fish exposed to natural photoperiodic regime , or earlier (between January and March) when the photoperiodic regime was altered (Breton et al., 1985). Our present data suggest that this increase in GtH is the reflection of short-term, high-amplitude, episodic elevations in the hormone level. The GtH pulses obse,rved in March were not synchronized among the individual fish and high GtH levels were recorded during no more than 2 hr. Thus, a point sampling of a group of fish undergoing early ovarian recrudescence (March), for the study of seasonal GtH variations, might shaw a mean elevated GtH level accompanied by high individual variation, as was actually the case (Breton et al., , 1985.
High-amplitude GtH pulses were observed in only 50% of the fish studied in March. This might be due to the short duration of the pulses and to then-low frequency (in most of the cases only one Gt pulse was observed during the 12-hr sampling period). However, the absence GtH pulses in half of the females shou also be considered together with the fact that in June no more such GtH pulses were observed, and with the findings of Breton et al. ( , 1985 concerning the transient nature of the spring GtH surge. Such's con-  sideration might indicate that the period during which the high-amplitude GtH pulses occur in each female is short, and that they characterize a precise given physiological state. A low degree of physiological synchronization among the studied fish might explain the differences found in their individual GtH profiles. According to Van den Hurk and Peute (1979) and Breton et al. ( , 1985, exogenous vitellogenesis in the rainbow trout (as observed histologically) first occurs in May, 1 to 2 months following the spring GtH rise (March). However, a recent study of the same species (Sumpter et al., 1984) has indicated that exogenous vitellogenesis can begin as early as March. Although Lambert et nl. (1978), Whitehead et al. (1978a, b), Scott et al. (1980), and Van Bohemen and Lambert (198 1) found very low and constant E,17B and vitellogenin levels in the female trout during spring, other studies have shown a slow increase in plasma E,17B levels starting in March  or even earlier (Fostier and Le Bail, unpublished data;Sumpter et al., 1984). Sumpter et al. (1984) reported an increase in circulating vitellogenin levels in the rainbow trout as early as January, and the beginning of its incorporation in late March. In the present study, the first signs of vitellogenin uptake were also histologically visible in the fish bled in March, in which the highamplitude GtH pulses were found. These considerations indicate that the GtH pulses might play a role in the induction of vitellogenin uptake by the follicles.
In fact, Breton and Derrien-Guimard (1983) have demonstrated that physiological GtH pulses stimulate vitellogenin incorporation into in vitro perifused oocytes of rainbow trout. However, our data, showing that E,17l3 levels in fish exhibiting high-amplitude GtH pulses were constant, do not necessarily mean that GtH pulses do not play a role in controlling the synthesis and/or secretion of E,17l3 from follicles at this time of the year. In a separate study, we have demonstrated that the administration of physiological GtH pulses to in vitro perifused ovarian fragments undergoing early recrudescence stimulated their E,17B output considerably (Zohar et al., 1982b;Zohar, Fostier, and Breton, unpublished results). In vivo, the GtH pulses might maintain constant E,17B levels over short periods.
The exact role of the pituitary in the control of the different phases of oogenesis has not yet been satisfactorily defined. Whereas the previtellogenic growth of oocytes seems to be independent of the pituitary (Vivien, 1939;Barr, 1963;Yamazaki, 1965), it has been shown in various teleost species that oogonial proliferation (Barr, 1963;Yamazaki, 1965) and vitellogenesis are pituitary-dependent processes (Barr, 1963;Yamazaki, 1965;Sundararaj et al., 1972;Khoo, 1979), the last one definitely being controlled by gonadotropin (Hoar et al., 1967;Hyder, 1972;Mackay, 1973;Upadhyay et al., 1978). On the basis of these observations, we cannot exclude the possibility that the high-amplitude GtH pulses found in March might also be related to the numerous oogonial mitoses and oocyte meioses, or to endogenous vitellogenesis, which were observed in the ovaries of the studied females.
In June, when most of the ovarian follicles were in the early stages of exogenous vitello.genesis, GtH (1-3 rig/ml) and E,17P (OS-l.5 rig/ml) levels were low and constant. In September-October, when exogenous vitellogenesis was in its advanced stages, plasma GtH levels fluctuated again. In most of the studied females, episodic, short-term GtH pulses of moderate amplitude were recorded, while basal GtH levels (2.4 rt 0.6 rig/ml) were only slightly higher than those recorded in June (1.8 * 0.4 ngiml). The appearance of GtH pulsatility during exogenous vitellogenesis was accompanied by a large increase of plasma E,17B, up to levels ranging from 6 to 30 rig/ml. At advanced exogenous vitello-genesis, individual E,17B profiles showed continuous and progressive variations superimposed on the abrupt GtH fluctuations. In contrast with the GtH profiles, a relative synchronization existed among the individual E,17B profiles. This was reflected by an average E,17B profile showing regular significant daily fluctuations (Fig. 6). E,17@ levels increased during the photophase and reached maxima toward and during early scotophase; a decrease later on might occur, which shonld be confirmed.
In the goldfish, Hontela and Peter (1978) found constant plasma GtH levels in "regressed" fish, in which oocytes undergoing endogenous vitellogenesis dominated the ovary. Daily fluctuations in GtH levels appeared later on, in "maturing" and in "mature" females. In the ovaries of these females, the proportion of oocytes undergoing advanced stages of vitellogenesis was high. In all groups, basal GtH levels were the same. Eased on these results, Hontela and Peter (1978) and Peter (1981) suggeste that the daily fluctuations in GtII levels might be more important for the stimulation and the maintenance of the ovarian activity than the progressive long-term changes in the concentrations of the hormone. On the basis of our present results, a similar hypothesis might be proposed for the female rainbow trout. Whereas previous studies suggested that the i~~p~rta~t increase in circulating E,17B duri nous vitellogenesis is accompani by a moderate increase or conti crease in the mean plasma GtH le constant concentrations of the (see introduction), our study su important change in the short-term GtI-I profiles during the s e processes. The constant levels of Gt haracterizing the beginning of exogenous vitellogenesis become pulsatile in its advanced stages, while E,17P increases drastically. The basal GtH levels remain low in both cases. This sug-ZOHAR, BRETON, AND FOSTIER gests that in the female rainbow trout the appearance of GtH pulsatility during exogenous vitellogenesis, rather than a change in the hormone absolute levels, is responsible for the marked E,17B increase. However, the rainbow trout model might be more complicated than the goldfish one. The results obtained by Hontela and Peter (1978), and confirmed by Hontela and Peter (1980a, b) and , indicate that there is a relative synchronization between the individual GtH profiles of different females. This fact results in the detection of a significant daily cycle in GtH levels when different groups of fish are sampled every 4 hr. Our present data demonstrate that in the female rainbow trout, the individual GtH profiles are not synchronized since the short-term increases in GtH levels occur at different times in different females. Thus, the GtH fluctuations can be detected only if individual fish are sampled repeatedly. Averaging the individual GtH profiles masked the existence of GtH pulsatility, and resulted in stable mean GtH levels over the 24-hr sampling period (Fig. 6). A similar situation has been described by Pan et al. (1980) in three carp species; when individual GtH profiles were recorded during the spawning season, short-term GtH pulses were observed. As in the case of the rainbow trout at advanced exogenous vitellogenesis, those profiles were relatively nonsynchronized among individual fish.
Although it is difficult to determine the frequency of GtH pulses during advanced exogenous vitellogenesis, we observed up to 5 GtH pulses within a 12-hr period when the bleeding interval was 1 hr. However, the number of pulses varied between the females, and in some of them GtH levels were quite constant. This might reflect a lack of total physiological synchronization among the studied females or might be the result of not bleeding frequently enough in relation to the duration of the pulses. In many of the profiles for which bleeding was done at l-hr intervals, a pulse included only one value of high GtH level. The pulses were longer in a few cases, and in one female (Fig. 5, No. 154) the increase in GtH levels lasted a few hours. This last example might reflect the initiation of an increase in the basal GtH levels, related to a possible initiation of the germinal vesicle migration (see Zohar et al., 1986). When the sampling interval was reduced to 30 min, most of the GtH pulses included more than one value of high GtH level. Considered together, all of the GtH profiles characterizing exogenous vitellogenesis indicate a pulse duration of 1 to 3 hr. Bleeding fish at even higher frequencies than that reported here is necessary for a precise analysis of the dynamics of the GtH pulses. Our present data do not demonstrate a clear circadian rhythmicity in GtH levels at advanced exogenous vitellogenesis.
However, they indicate a possible decrease in the frequency of the GtH pulses during the night, which should be further confirmed.
The relationship between the GtH and E,17B secretion patterns in the female trout at advanced exogenous vitellogenesis cannot be precisely determined on the basis of the present in viva study. They are considered in more detail elsewhere (Zohar, Fostier, and Breton, in preparation), taking into account the results of both in vivo and in vitro studies. However, the relationship between the GtH and E,17B secretion patterns differs from that described in most mammalian species, in which pulses are considered as basic signals at both the pituitary and the gonadal level. A pulsation of LH is immediately followed by a pulsation of a corresponding gonadal steroid in the mouse (Desjardins, 1981), the rabbit (Moor and Younglai, 1975), the goat (Muduuli et al., 1979), the dog (De Palatis et al., 1978), the ewe (Baird, 1978), the ram (Lincoln, 1976), the rhesus monkey (Steiner et al., 1980), and man (Backstrom et al., 1982). In the female rainbow trout, however, the short-term GtH pulses are accompanied by regular E,17P variations of longer duration. The same phenomenon was also observed in vitro (Zohar et al., 1982b and unpublished data). The greater abundance of GtH pulses during the day may indicate the importance of their frequency in the regulation of the progressive increase in E,17P levels during the photophase, as well as in the possible decrease of these levels later on.
Recently, the physiological function of fish gonadotropin(s) in relation to vitellogenesis has been a matter of discussion (e.g., Idler, 1982;Idler and Ng, 1983;Burzawa-Gerard, 1982). In this discussion, references have been made to results which are related to the circulating levels of the glycoproteic GtH. Our present study is the first to demonstrate that in the female rainbow trout, changes in the short-term secretion pattern of GtH occur from early ovarian recrudescence and throughout vitellogenesis, in addition to the seasonal evolution of the hormone level. Such short-term secretion patterns should be considered when the glycoproteic GtH role in vitellogenesis is discussed. Zohar, Y., Breton, B., and Billard, R. (1982a). Shortterm profiles of plasma gonadotropin levels in the female rainbow trout throughout the reproductive cycle. Gen. Comp. Endocrinol.