Role of Diameter Differences among Follicles in Selection of a Future Dominant Follicle in Mares'

Transcription

1 BIOLOGY OF RPRODUCTION 57, (1997) Role of Diameter Differences among Follicles in Selection of a Future Dominant Follicle in Mares'.L. Gastal, M.O. Gastal, D.R. Bergfelt, and O.J. Ginther 2 Animal Health and Biomedical Sciences, University of Wisconsin-Madison, Madison, Wisconsin 5376 ABSTRACT Follicles - 5 mm were ablated in pony mares by a transvaginal ultrasound-guided technique on Day 1 (ovulation = Day ). Follicle emergence (at 15 mm, experiment 1; at 6 mm, experiment 2) and development of the new wave was monitored by transrectal ultrasound. Deviation was defined as the beginning of a marked difference in growth rates between the two largest follicles. In experiment 1, mares were grouped (n = 4 per group) into controls, ablation-controls (ablations at Day 1 only), and a two-follicle model (periodic ablation sessions so that only the two largest follicles developed). There were no significant indications that the two-follicle model altered follicle diameters, growth rates, or time intervals of the two retained follicles at or between events (follicle emergence, deviation, and ovulation). In experiment 2, the two-follicle model (n = 14) was used for follicle and hormonal characterization and hypothesis testing, without the tedious and error-prone necessity for tracking many (e.g., 2) individual follicles. The future dominant follicle emerged a mean of 1 day earlier (p <.8) than the future subordinate follicle, the growth rates for the two follicles between emergence and deviation (6 days later) did not differ, and the dominant follicle was larger at the beginning of deviation ( mm versus 19.6 ±.9 mm; p <.1). Mean FSH and LH concentrations increased (p <.5) concomitantly from emergence of the future dominant follicle and peaked 3 days later when the follicle was a mean of 13 mm. Thereafter, the two hormones disassociated until ovulation: FSH decreased and LH increased. Results supported the hypothesis that the future dominant follicle has an early size advantage over future subordinate follicles and indicated that the advantage was present as early as 6 days before deviation. INTRODUCTION Tracking of individual follicles using ultrasonography has established that follicular development occurs in waves in mares [1, 2] as in cattle [3]. Most mares have one major wave that gives origin to the ovulatory follicle. However, other mares may have two major waves during an estrous cycle. In major waves, the dominant follicle grows to a preovulatory diameter - 3 mm, whereas the largest subordinate follicle regresses after reaching approximately 22 mm. Furthermore, minor waves sometimes develop in which the largest follicle reaches reported means of mm and no follicle establishes dominance [2, 4-6]. Growing follicles of a major wave and the regressing follicles of a preceding major or minor wave can intermingle, thereby Accepted July 15, Received April 2, 'Supported by the College of Agricultural and Life Sciences, University of Wisconsin-Madison and by quiculture, Inc., Cross Plains, WI. This work was presented in part at the 23rd Annual Conference of the International mbryo Transfer Society, Nice, France, L.G. and M.O.G. are supported by the Federal University of Vi;osa and by a CAPS scholarship, Brazil. 2Correspondence: Animal Health and Biomedical Sciences, University of Wisconsin-Madison, 1655 Linden Drive, Madison, WI FAX: (68) ; ojg@ahabs.wisc.edu 132 complicating the study of individual follicles of a wave. In addition, follicles in a given wave can be numerous. These factors can lead to crowding of follicles, and tracking individual follicles < 15 mm by ultrasound can be laborintensive, tedious, and error-prone. Accurate identification of individual small follicles (e.g., 6-14 mm) from examination to examination would permit in-depth investigation of wave characteristics. Thus, a simplified follicle model in this species would aid in the study of some of the fundamental aspects of follicular wave dynamics. The mechanism involved in selection of one follicle of a follicular wave to become dominant in monovular species while other follicles become subordinate and regress is under intense investigation [3]. Wave emergence has been defined as the day or time a future dominant follicle reaches a specified size in mares [2] and cattle [3]. Deviation is a term that has been used in cattle to describe the beginning of a pronounced difference in growth rates between the two largest follicles [3]. In cattle, the future dominant follicle was detected as a 3-mm follicle at a mean of 6 h earlier than for the future largest subordinate follicle [3]. Furthermore, mean diameter of the dominant follicle at the beginning of deviation was larger than for other follicles of the wave. These results supported the hypothesis that the future dominant follicle had an early developmental advantage and suggested that the follicle selected to become dominant, as manifested by deviation, was the first follicle to develop to a decisive stage. This size-advantage hypothesis has not been tested in mares. Surges in circulating concentrations of FSH and LH are more synchronous during the luteal phase than during the follicular phase in mares [7]. Surges of FSH, but not LH, have been studied in relation to emergence of major and minor waves [2, 4, 6, 8]. The association and the disassociation between FSH and LH have not been studied in relation to follicular wave dynamics in this species. The objectives of the present study in mares were as follows: 1) to develop and validate a simplified model for tracking follicles by allowing the two largest follicles of a wave to develop while other follicles are ablated (experiment 1); 2) to characterize follicular and hormonal dynamics of the two-follicle model (experiment 2); and 3) to test the hypothesis that the future dominant follicle has an early size advantage over other follicles of the wave (experiment 2). MATRIALS AND MTHODS Animals and Ultrasonography Nonlactating pony mares, 4-2 yr and kg, were used from March to April (experiment 1) and June to July (experiment 2). The mares were kept in outdoor paddocks under artificial light (15 h per day) from November to March preceding the study. They were maintained on alfalfa/grass hay and had access to water and trace-mineralized salt. The ultrasound scanner was equipped with a

2 DOMINANT AND SUBORDINAT FOLLICLS IN MARS mHz linear-array intrarectal transducer (Tokyo Keiki LS3; Products Group International, Lyons, CO) for sequential examinations of the reproductive tract. A scanner (Aloka SSD-5V; Aloka, Wallingford, CT) equipped with a 5-mHz convex-array transvaginal transducer (Aloka UST974V-5) was used for follicle ablations. Ultrasonic examinations of the ovaries and uterus were done daily or every other day as described [9], and mares were not used if they had indications of ovarian or uterine abnormalities. All mares were determined to have entered the ovulatory season by ultrasonic examinations for detection of ovulation and the formation of a corpus luteum [9]. Before the start of the study, mares were synchronized with prostaglandin F 2, (Lutalyse; Pharmacia & Upjohn Co., Kalamazoo, MI). Follicles > 3 mm in diameter were monitored daily until ovulation (Day ). The experiments were started on Day 9, and examinations of the ovaries were done by a single operator. The average of height and width of the antrum at the maximal area was used as a measure of follicle diameter. Follicle diameters were determined from a single image unless otherwise indicated. mergence of a follicle was defined as occurring on the day the follicle was 15 mm in experiment 1 or 6 mm in experiment 2. The dominant follicle (one that grew to a preovulatory size - 3 mm) and largest subordinate follicle (one that grew to a moderate size and regressed) were chosen retrospectively by the maximum attained diameter. The day of deviation between the future dominant and largest subordinate follicles was defined as the day at the beginning of the greatest difference in growth rates (diameter changes between adjacent examinations) between the two follicles at or before the day when the largest subordinate follicle reached maximum diameter or an apparent diameter plateau. The relative location of follicles, corpus luteum, and follicle ablation sites (echoic areas) were used as references in identifying and tracking individual follicles. The diameter of the corpus luteum was evaluated daily as previously described [9]. To minimize the potential of obscuring the results of the characterization studies and the test of the hypothesis, preplanned criteria were established for removing mares from the experiments. The criteria are given in the results section of each experiment. Follicle Ablation Mares were held in a padded squeeze chute to prevent excessive movement. Sedation and analgesia were induced with xylazine (Rompum, 1 mg/kg i.v.; Miles Inc., Shawnee Mission, KS) and butorphanol tartrate (Torbugesic,.5 mg/kg i.v.; Fort Dodge Laboratories, Fort Dodge, IA) in experiment 1 or detomidine hydrochloride (Dormosedan,.2-.4 mg/kg i.v.; Pfizer Animal Health, West Chester, PA) and butorphanol tartrate in experiment 2. In experiment 1, but not in experiment 2, caudal epidural anesthesia was induced with 6-8 ml of 2% lidocaine hydrochloride (Phoenix Pharmaceutical, Inc., St. Joseph, MO). In experiment 2, rectal relaxation was induced with hyoscine N-butyl bromide (Buscopan,.2 mg/kg i.v.; Sigma Chemical Co., St. Louis, MO). A tail bandage was applied, and the perineal area was aseptically prepared for follicle ablation. Follicle ablation was done transvaginally by ultrasoundguided aspiration of follicular contents similar to that described in cattle [1]. A 17-gauge needle was used to puncture individual follicles. Follicular contents were removed using a vacuum pump (25-3 mm Hg). No attempts were made to determine the presence of an oocyte or number of granulosa cells within aspirates. Follicle ablation was defined by collapse of the antral follicle after evacuation of follicular contents. The ovaries were examined 2 h after every ablation to ensure that the targeted follicles were completely ablated. At the next ablation session, ablated follicles that refilled were ablated again. xperiment 1 Mares were randomly assigned to 1 of 3 groups on Day 9: 1) control (no follicle ablation, n = 6), 2) ablation control (n = 6), and 3) two-follicle model (n = 6). Beginning on Day 9, all follicles - 5 mm were counted daily. An attempt was made to track individual follicles - 8 mm in all groups until ablation, regression, or ovulation; however, tracking follicles of 8-14 mm often was difficult, especially in the control group. Therefore, 15 mm was adopted as the minimal diameter for tracking purposes and to establish the day of emergence of a follicle. In the two ablation groups (ablation control and two-follicle model), all follicles - 8 mm were ablated on Day 1. In the two-follicle model, a second ablation was done when the largest follicle reached 16 mm; all follicles 8 mm were ablated except the two largest. Similarly, subsequent ablations were done when follicles, other than the two largest, reached 1 mm, and ablation sessions continued until the largest follicle reached 25 mm. This was done to simplify follicle tracking at the approximate time of deviation. xamples of follicular profiles before and after ablation for one mare in each group are shown (Fig. 1). xperiment 2 Mares were assigned to the two-follicle model (n = 21) on Day 9. Follicles - 5 mm were measured daily until the largest follicle reached 15 mm; thereafter, the height and width of the two largest follicles were measured with three new images during a scanning session. The average of the six measurements per follicle was used as the follicle diameter for each day until the next ovulation. The ablation and tracking protocols were similar to those in the corresponding group in experiment 1, except that all follicles _ 5 mm were ablated on Day 1. This was done to study follicle growth at smaller diameters. In addition, subsequent ablations were done when the largest follicle reached 15 mm; all follicles 5 mm were ablated except the two largest. Periodic ablations continued until the largest follicle reached 25 mm as for experiment 1. Characterization of follicular dynamics was similar to experiment 1 except that the day of follicle emergence was assigned to the day before a follicle exceeded 6 mm. Blood samples were collected daily beginning on Day 1 and ending on the day of the next ovulation. Samples were collected by jugular venipuncture into heparinized tubes and held for 1-2 h at 4 C until sedimentation. Plasma was decanted and placed in vials for cold storage (-2 C) until assay. Plasma concentrations of FSH [11], LH [12], and progesterone [13] were determined using RIAs validated in this laboratory for this species. For the FSH, LH, and progesterone assays, the within- and between-assay coefficients of variation were 7.4% and 3.9%, 9.8% and 2.9%, and 11.2% and 4.2%, respectively, as determined over two assays each for FSH, LH, and progesterone.

3 1322 GASTAL T AL Control S ov 32 - ) Cu ) 12 - a)._d ion I - OI Ablation OV 24. : * S S * S S Two-follicle model Ablatecd follicles$ (n=22) I * : (n=6) * * * * ::: *. OV ; Days after ovulation FIG. 1. xamples of follicular profiles for 1 mare in each group. Profiles of the dominant (solid line with dots), largest subordinate (dashed line with opened dots), and smaller subordinate (dotted line with solid dots) follicles are shown. The unconnected solid dots represent follicles that were not tracked. In some instances, a solid black dot of untracked follicles represents more than one follicle. Day of emergence of the future dominant follicle at - 15 mm (solid arrow) and day of deviation (dashed arrow) are indicated for each mare. OV, ovulation. xperiment 1. Statistical Analyses In experiments 1 and 2, analysis of variance was used for determining an effect of group for single-point measurements, and paired or two-sample t-tests applied to the ranks [14] were used to compare various characteristics between the dominant and largest subordinate follicles. In experiment 1, split-plot analysis of variance was used for de L Largest subordinate follicle Two-follicle model Days after dominant follicle reached 15 mm FIG. 2. Mean ( SM) diameters of the dominant and largest subordinate follicles normalized to the day the dominant follicle reached 15 mm. There was an effect of day (p <.1) but no significant effect of group or group-by-day interaction for each end point. n = 4 mares per group. xperiment 1. termining main effects of group and day, and the interaction of group-by-day for sequential follicular data. In experiment 2, analysis of variance was used to determine an effect of day for sequential follicular and hormonal data. Variation due to the sequential nature of the data was accounted for by using mare (group) as the whole-plot error term to test the effect of group in experiment 1, or by including mare in the model and using the mare-by-day interaction as the error term to test the effect of day in experiment 2. In experiments 1 and 2, if a significant effect of group or day or an interaction of group-by-day was indicated, Duncan's Multiple Range tests were used to locate mean differences among groups, among groups within days, and among days within groups. Significance was defined as p <.5. RSULTS xperiment 1 nab si Ir Six mares were removed from the experiment for the following reasons: largest subordinate follicle did not exceed 15 mm (failure to develop may not have involved the deviation mechanism), 4 mares; double ovulations (codominant follicles), 1 mare; and largest follicle regressed in the presence of a maintained corpus luteum (anovulatory wave), 1 mare. Removed mares were equally distributed over treatments (2 per group). Thus, a total of 4 mares per group were available for statistical analyses. On Day 1, mean ( SM) total number of follicles 2 5 mm was similar among control, ablation control, and the two-follicle model groups (19.5 ± 2.1, 18.8 ± 2.2, and follicles, respectively). However, on each of Days there were more (p <.1) follicles in the control group than in each of the ablation groups. The range of

4 DOMINANT AND SUBORDINAT FOLLICLS IN MARS 1323 TABL 1. Mean ( SM) intervals, follicle diameter, and growth rates in control and follicle-ablated mares. xperiment 1 xperiment 2 Ablation Two-follicle All Two-follicle Control control model a b groups modelab nd point (n = 4) (n = 4) (n = 4) (n = 12) (n = 14) Intervals (days) from: Day 1 c to emergenced of follicle Dominant follicle X x Largest subordinate follicle Y 2.7 ±.4Y mergence of follicle to deviatione Dominant follicle X x Largest subordinate follicle ± ± Y Y Deviation to ovulation ± Day 1 to ovulation ± Follicle diameters (mm): On day of emergence of dominant follicle Dominant follicle ± ± x X Largest subordinate follicle ± ± ±.8 Y <5.2Yh At the beginning of deviation Dominant follicle ± ± x X Largest subordinate follicle ±.9 2. ± Y Y On day before ovulation of dominant follicle Dominant follicle ± X x Largest subordinate follicle ± ± ± 1.Y Y Growth rates (mm/day) from: mergence of follicle to deviation Dominant follicle ± Largest subordinate follicle Deviation to ovulation of dominant follicle Dominant follicle 2.1 ±.2 f 2.9 ±.1g 2.7 ±.3g 2.5 ±.2 X x Largest subordinate follicle -.7 _ ± Y Y a All follicles 2 8 mm (experiment 1) or - 5 mm (experiment 2) ablated on Day 1, postovulation. bwhen the largest follicle of the post-ablation wave reached 16 mm (experiment 1) or 15 mm (experiment 2) all follicles were ablated, except the 2 largest. c Day 1 = 1 days after ovulation. d mergence of a follicle defined as the day it reached 15 mm (experiment 1) or 6 mm (experiment 2). e Deviation = day at beginning of greatest difference in growth rates between the 2 largest follicles. f,g Means within rows for the 3 groups with no common superscript are different (p <.5). h Value unknown in 3 mares because the subordinate follicle was not detectable on day of emergence of dominant follicle; data analyzed by a ranking test. XY Paired means in experiment 1 (all groups) or experiment 2 with a different superscript are different (p <.3 to p <.1). mean follicle numbers on Days 11 to 16 was as follows: controls, ; ablation controls, ; and two-follicle model, There were no significant differences between the two ablation groups. Number of follicles from 1 mare in each group is illustrated (Fig. 1). There was no effect of group or a group-by-day interaction for diameters of the future dominant and largest subordinate follicles during the 6 days after the dominant follicle reached 15 mm (Fig. 2); data were truncated at 6 days because of ovulations. Intervals from one point of reference to another for dominant and largest subordinate follicles were not significantly different among groups (Table 1). The mean interovulatory interval, as reflected by the interval from Day 1 to ovulation, was similar among groups. There were no significant differences involving follicle diameter and growth rates, except that the diameter growth rate of the dominant follicle after deviation was higher (p <.5) in the ablation control group than in the control group. Although there were differences among groups for 1 of 7 end points involving the future dominant and largest subordinate follicles (Table 1), data were combined across groups to increase the number of observations for comparing the dominant and largest subordinate follicles within each end point. There were significant differences between the dominant and largest subordinate follicles for each end point in Table 1 except that the diameter growth rate from emergence to deviation was not different between the two follicles. The dominant follicle emerged earlier (p <.1) than the largest subordinate follicle. The dominant follicle was larger than the largest subordinate follicle on the day of emergence of the dominant follicle (p <.9) and on the day of deviation (p <.2). xperiment 2 Technical aspects of the two-follicle model for the 21 mares were as follows: a total of 46 ablation sessions were done with a mean of 2.2 sessions per mare, taking approximately 15 min per session; a mean of 8.5 follicles were ablated per session with a total of 19 ablated follicles per mare; mean number of sessions per mare and number of ablated follicles 5 mm per session progressively decreased over the first (15.2 ± 1.4 follicles; n = 21), second (3.2 ±.6 follicles; n = 2), and third (1.2 ±.7 follicles; n = 5) ablation sessions. Re-ablation of follicles was done because of refilling in 43 of 2 (21.5%) follicles, and rec-

5 1324 GASTAL T AL. (c) b _ I -o 8- C - 6- (/) 4-2- AN 4U O a (b).. I. I I I I I Dominant follicle,./ 9 Subordinate,4f. " follicle Mares with follicle deviation on indicated days (1) (2) (6) (2) (2) (1) Days from emergence of dominant follicle (d),! I.. I Days from deviation between dominant and subordinate follicles FIG. 3. Mean (+ SM) day-to-day diameter of the dominant and subordinate follicles and circulating concentrations of FSH, LH, and progesterone. The data are normalized to the day of emergence of the dominant follicle (a, b) and the day of deviation (c, d). Normalizing the data to the day of deviation produced a sharper diameter departure between follicles (d vs. b) but did not indicate any distinctive hormonal changes on the day of deviation (c). There was a significant effect of day (p <.1) for each end point. * First significant (p <.5) increase or decrease in concentrations of each hormone. tal irritation, colic, infection, or palpable ovarian adhesions were not observed. Seven mares were removed from the experiment for the following reasons: subordinate follicle did not exceed 15 mm, 1 mare; day of deviation was indistinct, 2 mares; largest follicle regressed in the presence of a maintained corpus luteum, 2 mares; and unsuccessful follicle ablation, 2 mares. Thus, a total of 14 mares were available for statistical analyses. Intervals from one point of reference to another, and follicle diameters and growth rates for the future dominant and subordinate follicles are shown (Table 1). Similar to experiment 1, the only end point that was not significantly different between the two follicles was diameter growth rate from emergence of each follicle to deviation. The dominant follicle emerged earlier (p <.8) and was larger than the subordinate follicle on the day of emergence of the dominant follicle (p <.2) and on the day of deviation (p <.1). Mean day-to-day changes in diameter of the future dominant and subordinate follicles, using the day of emergence of the future dominant follicle at 6 mm and the day of deviation as reference points, are shown (Fig. 3). Data were truncated to the days when observations were available for all mares. The mean day of deviation was days after emergence of the dominant follicle or on Day of a mean interovulatory interval of 24 days. Changes in concentrations of FSH, LH, and progesterone are shown normalized to the two reference points (Fig. 3). Data profiles of hormones, follicles, and corpora lutea for 3 selected mares are also shown (Fig. 4). Included are data for a mare that had an 5 4 a a 3 2 1

6 DOMINANT AND SUBORDINAT FOLLICLS IN MARS 1325 ' X- IX> Io U- Ub CU 'a azo ; Days after ovulation Mare C... le * IWSU S i,i i I I 8 Anovulatory o I o 12 co r- C 2 C I FIG. 4. Day-to-day diameters of the largest and second largest follicles and corpus luteum, and circulating concentrations of FSH, LH, and progesterone for 3 selected mares. Mares were chosen to illustrate the temporal associations between deviation and a transient elevation in LH concentrations (mare A; 57% of the mares) or an LH plateau (mare B; 21% of the mares). Mares A and B developed a prominent dominant follicle (largest follicle). Mare C is an example of 1 of 2 mares that were removed from the experiment because of partial regression of the corpus luteum and the apparent lack of development of dominance by the largest follicle. anovulatory wave with a persistent corpus luteum; the mare was removed from the data analyses. DISCUSSION The removal of 13 of 39 (33%) mares from the data analyses was large but was done according to the protocol. In these initial studies of the two-follicle model, it was deemed appropriate to remove mares with an obscure deviation mechanism (possible failure of the two follicles to partition into dominant and subordinate follicles). Occasionally (1 of 39; 3%), both follicles became dominant, as indicated by double ovulations, resulting in removal of the mare. The double ovulation rate was similar to a reported rate of 2% [15]. In other removed mares (13%), the second largest follicle grew to < 15 mm. That is, the follicle may have regressed because of poor development rather than in response to the deviation mechanism. The planned removal of mares when the wave did not produce an ovulatory follicle was prudent because of the likelihood that the largest follicle would be too small to express dominance (no deviation mechanism). In mares removed for this reason (8%), the corpus luteum regressed incompletely at the expected time, resulting in a prolonged interovulatory interval (39-42 days compared to a normal mean of 24 days) [15]. A partially maintained corpus luteum in a mare with a minor wave (no apparent dominance) is illustrated (mare C, Fig. 4). Prolonged maintenance of the corpus luteum occurs occasionally during the estrous cycle in mares [16], and failure of a follicular wave to contribute a dominant follicle is common during early pregnancy in which the corpus luteum is maintained [6, 8]. Sequential, ultrasonographic studies of follicular waves l ; 3 in mares have used grouping of follicles according to diameters [17, 18], measurement of diameter of the two largest follicles [17-19], determination of wave characteristics by statistical changes in diameters [2], and grouping of follicles into tiers of the 6 largest follicles per tier for many tiers [4]. These methods did not require maintaining identity of individual follicles. Removal of the preovulatory follicle and studying the interval to development of another ovulatory-sized follicle [1, 21], and determining when large follicles no longer respond to stimulation [22] have also been used to study follicular development. In addition, tracking individual follicles from examination to examination has been done until the follicles, in retrospect, were approximately 15 mm [1, 2, 6] or by selecting mares with few follicles per ovary [23]. All of the above techniques advanced our knowledge of follicular dynamics in mares but have not allowed detailed study of individual small follicles (e.g., 6 mm). Furthermore, tracking a full complement of follicles in this species is time-consuming, stressful to the mare and operator, and error-prone. For these reasons, the two-follicle model was developed. In experiment 1, ablation of all follicles - 8 mm on Day 1 was an aid in studying the resulting newly emerging wave because regressing follicles of a previous wave would have obscured the identity of follicles of the new wave. The ablations on Day 1 reduced the number of follicles to be monitored on Days by approximately 5%. Thus, in retrospect, follicle tracking could have begun at 6 mm, rather than 15 mm. Subsequent ablations after the follicles of the new wave reached 16 mm in the two-follicle model was an additional aid. The two follicles were easier to track, beginning well before deviation, required only a 5

7 1326 GASTAL T AL , 2 a) Days after larger follicle reached 6 mm FIG. 5. An interpretation of results for the two-follicle model on the nature of the deviation mechanism. The larger follicle (future dominant) has a size advantage and is the first to reach a critical stage (black bar) associated with a mechanism that inhibits the smaller follicle (future subordinate) before it can reach a similar stage (white bar). few minutes per examination, and probably minimized error in identity. In contrast, attempts at follicle tracking in the control mares required 15-3 min per mare. There were no significant differences in experiment 1 that would indicate that the two-follicle model altered time intervals, diameter growth profiles, or diameters of the two largest follicles at or between various times (emergence, deviation, ovulation). The only significant difference among the three groups was a slower diameter growth rate of the dominant follicle between deviation and ovulation in controls compared to ablation controls; the reason for this result is not known and requires confirmation. The corresponding diameter growth rate in the two-follicle model did not differ from that of either of the control groups. Although there were no significant differences among groups between the day of emergence and the day of deviation, it is emphasized that the number of mares per group (n = 4) was small. In addition, plasma samples were not obtained for hormonal comparisons among groups. Therefore, we retain reservations about using the model for studying the two largest follicles in intact mares. However, we believe the model is suitable for initial work and will give direction to later studies in intact mares. Characteristics of the two follicles were similar between experiments for the two-follicle model, considering that emergence of a follicle was based on 15 mm in experiment 1 and 6 mm in experiment 2. In xperiment 2, the beginning of deviation in diameter growth rates occurred on mean Day 18 or 6 days after emergence of the dominant follicle and 7 days before ovulation. An interval of 7 days between deviation and ovulation agrees with previous reports involving transrectal palpation and ultrasound studies [9,15]. The dominant follicle was detected at emergence (6 mm) approximately 1 day earlier than the subordinate follicle. Similarly, on the day of its emergence, the future dominant follicle was larger than the future subordinate follicle. There was no difference in diameter growth rates between the two follicles from the time of emergence of a follicle at 6 mm until the defined day of deviation. On the day of deviation, the diameters of the future dominant and subordinate follicles averaged 23 and 2 mm, respectively. 1 It follows that the dominant follicle, on average, maintained its larger diameter from emergence to deviation. Thus, the hypothesis of an early size advantage of the future dominant follicle over other follicles of the wave was supported. Although the hypothesis was supported by the differences in means, there were exceptions in individual mares. In 1 of 14 (7%) mares, the future subordinate follicle was larger than the future dominant follicle on the day the dominant follicle was 6 mm. Five days later, the future dominant follicle was larger, and it thereafter maintained its size advantage. In all 14 mares, each of the two largest follicles maintained the position of being either the largest or second largest between the time the future dominant follicle was 6 mm and the second ablation session at 15 mm (approximately, 4 days); that is, there were no instances of a smaller follicle becoming one of the two largest. In 2 mares, the future subordinate follicle was larger on the day of the second ablation session when the largest follicle of the wave was 15 mm. In both of these mares, the smaller retained follicle became larger by the time of deviation. A change in position of the two follicles was more likely to occur when the diameter difference between follicles was small (1 or 1.5 mm). Although these exceptions could be attributed to measuring errors, it seems more likely that the two largest follicles occasionally can change positions between the day of emergence and the day of deviation; similar exceptions have been reported in cattle [3]. In all 14 mares, however, the follicle that was larger at the beginning of deviation became the dominant follicle. In a recent report in cattle [3], the future dominant follicle began as a 3-mm follicle a mean of 6 h earlier than the future largest subordinate follicle. Hence, ponies and cattle appear similar in that eventual designation of the dominant follicle at deviation involves a size advantage; the size advantage was equivalent to a mean of 6 h in cattle and I day in ponies. Several studies in cattle have indicated that removal of the largest follicle before or soon after deviation allows the second largest follicle to become dominant (reviewed in [3]). These results indicate that final designation of the dominant follicle does not occur before deviation. Similar studies have not been done in mares; however, the findings in cattle and the present findings in ponies warrant focusing on the time of deviation in studying the mechanism of selection of a dominant follicle. Our interpretation of the results for the two-follicle model is that the follicle destined to become dominant was the first to reach a critical stage associated with deviation. The deviation mechanism inhibited the other follicle before it reached a similar stage (Fig. 5). This interpretation implies that the event or condition responsible for deviation results in a positive or permissive effect on the selected larger follicle and a negative effect on the smaller follicle. If a specific signal is responsible for deviation, it seems likely that the breadth of the signal would be less than the time represented by the difference in diameter between the two follicles (equivalent to approximately 1 day). The role of FSH and LH in the deviation mechanism is unknown. Concentrations of FSH increased after the day the future dominant follicle was 6 mm, peaked 3 days later, and then declined. On average, these changes in FSH concentrations after ablation represent the FSH surge that stimulated emergence of the new follicular wave. The association between an FSH surge and the emergence of a follicular wave in mares has been previously shown for major and minor waves of the estrous cycle [2] and early pregnancy [6, 8]. The present data extend those findings by

8 DOMINANT AND SUBORDINAT FOLLICLS IN MARS 1327 demonstrating the association between the hormonal surge and follicles < 15 mm and show that the peak of the surge occurred, on average, when the largest follicle was 13 mm. This compares with the peak of the FSH surge occurring when the follicles are 4 mm in cattle [3]. After the FSH peak, the concentrations declined, and they reached nearminimal levels a few days later at the time of follicle deviation. In cattle, FSH concentrations reached a nadir at deviation [3]. Until the day of the FSH peak, which occurred during the progesterone decline, FSH and LH were closely associated; they then became profoundly disassociated until ovulation. This is consistent with a report [7] that fluctuations in concentrations of these two hormones are highly synchronous during the luteal phase and less synchronous during the follicular phase. The concomitant FSH and LH surges that occur before ovulation in cattle [24] do not occur in mares [15]. In the present data, there was an interval of 3-4 days between the beginning of the FSH/LH disassociation and the occurrence of deviation. Deviation occurred during high LH and low FSH concentrations, and either or both hormonal conditions may be necessary for deviation to occur. Although not preplanned, observational changes in LH concentration in each mare encompassing follicle deviation are presented for discussion. In 11 of 14 mares (79%), the day of deviation was encompassed by either an apparent transient LH elevation (8 mares; Fig. 4, mare A) or an apparent plateau (3 mares; mare B). In the remaining 3 mares (21%), LH increased progressively over 6-8 days before ovulation with no transient elevation or plateau. The transiently high or plateaued concentrations of LH encompassing deviation, as determined by daily blood sampling, did not meet the requirements of the deviation signal postulated in Figure 5, since the breadth of the change in LH was wider than 1 day in both mean and individual profiles. If there is a signal for deviation and if the signal is gonadotropic, future studies may require frequent blood sampling (e.g., every 15 min). In conclusion, our current working hypothesis for the two-follicle model is that the follicle destined to become dominant has a size advantage over the other follicle. The two follicles grow at a similar rate until the day of deviation, when the future dominant follicle continues to grow and the future subordinate follicle begins to regress. The follicle destined to become dominant is the first to reach a critical stage of approximately 23 mm and, at that time, is involved in a deviation mechanism that inhibits the other follicle before it reaches a similar stage, an average of 1 day later. The specific mechanism of inhibition and whether it is direct or involves the pituitary is not known. ACKNOWLDGMNTS The authors thank Lisa Kulick and Travis Clary for graphics and computer assistance, and the National Hormone Pituitary Program for antihuman FSH-6. RFRNCS 1. Sirois J, Ball BA, Fortune J. Patterns of growth and regression of ovarian follicles during the oestrous cycle and after hemiovariectomy in mares. quine Vet J 1989; 8(suppl): Bergfelt DR, Ginther OJ. Relationship between FSH surges and follicular waves during the estrous cycle in mares. Theriogenology 1993; 39: Ginther OJ, Wiltbank MC, Fricke PM, Gibbons JR, Kot K. Selection of the dominant follicle in cattle. Biol Reprod 1996; 55: Ginther OJ, Bergfelt DR. Growth of small follicles and concentrations of FSH during the equine oestrous cycle. J Reprod Fertil 1993; 99: Ginther OJ. Major and minor follicular waves during the equine estrous cycle. J quine Vet Sci 1993; 13: Bergfelt DR, Ginther OJ. Relationship between circulating concentrations of FSH and follicular waves during early pregnancy in mares. J quine Vet Sci 1992; 12: Bergfelt DR, Ginther OJ. Synchronous fluctuations of LH and FSH in plasma samples collected daily during the estrous cycle in mares. Theriogenology 1993; 4: Ginther OJ, Bergfelt DR. Associations between FSH concentrations and major and minor follicular waves in pregnant mares. Theriogenology 1992; 38: Ginther OJ. Ultrasonic Imaging and Animal Reproduction: Horses. Book 2. Cross Plains, WI: quiservices Publishing; 1995: Bergfelt DR, Lightfoot KC, Adams GP Ovarian synchronization following ultrasound-guided transvaginal follicle ablation in heifers. Theriogenology 1994; 42: Freedman LJ, Garcia MC, Ginther OJ. Influence of photoperiod and ovaries on seasonal reproductive activity in mares. Biol Reprod 1979; 2: Whitmore HL, Wentworth BC, Ginther OJ. Circulating concentrations of luteinizing hormone during estrous cycle of mares as determined by radioimmunoassay. Am J Vet Res 1973; 34: Bergfelt DR, Ginther OJ. Follicular populations following inhibition of follicle stimulating hormone with equine follicular fluid during early pregnancy in the mare. Theriogenology 1986; 26: SAS. SAS Procedures Guide, version 6, 3rd ed. Cary, NC: Statistical Analysis System Institute, Inc.; 199: Ginther OJ. Reproductive Biology of the Mare: Basic and Applied Aspects, 2nd ed. Cross Plains, WI: quiservices; 1992: Ginther OJ. Prolonged luteal activity in mares-a semantic quagmire. quine Vet J 199; 22: Ginther OJ, Pierson RA. Ultrasonic anatomy of equine ovaries. Theriogenology 1984; 21: Pierson RA, Ginther OJ. Follicular populations during the estrous cycle of the mare. Anim Reprod Sci 1987; 14: Palmer, Driancourt MA. Use of ultrasonic echography in equine gynecology. Theriogenology 198; 13: Ginther OJ, Bergfelt DR. Ultrasonic characterization of follicular waves in mares without maintaining identity of individual follicles. J quine Vet Sci 1992; 12: Driancourt MA, Palmer. Time of ovarian follicular recruitment in cyclic mares. Theriogenology 1984; 21: Pierson RA, Ginther OJ. Ovarian follicular response of mares to an equine pituitary extract after suppression of follicular development. Anim Reprod Sci 199; 22: Palmer. New results on follicular growth and ovulation in the mare. In: Roche JF, O'Callaghan D (eds.), Follicular Growth and Ovulation Rate in Farm Animals. Dublin: Martinus Nijhoff Publishers; 1987: Hansel W, Convey M. Physiology of the estrous cycle. J Anim Sci 1983; 57(suppl 2):