All procedures were performed by trained personnel in strict accordance with Spanish guidelines for the protection of experimental animals (RD 53/2013; BOE, 2013), and in agreement with European Union Directive 86/609. The study was conducted at the University of Huelva experimental farm (37º 20ºN, 6º 54’ W), which meets the requirements of the European Community Commission for Scientific Procedure Establishments (2010/63; OJEU, 2010).

The animals examined were seven Blanca Andaluza bucks (8 months old at the start of the experiment), previously trained to mount a teaser doe and to ejaculate into an artificial vagina. These animals were fed daily with barley straw (ad libitum), lucerne hay and a commercial concentrate, according to their body weight and in agreement with INRA standards (Morand-Fehr & Sauvant, 1988). All animals had free access to water and mineral blocks containing trace elements and vitamins.

Data collection began on 14^th November 2012 and ended on 9^th July 2014. However, since the bucks were young at the start of the experiment, only data from the last year (July 2013 to July 2014) were used in analyses. The experiment was designed to determine the effect of season on reproductive status and sperm variables. Summer, autumn, winter and spring were defined as the periods between June 21^st and September 22^nd, September 23^rd and December 20^th, December 21^st and March 20^th, March 21^st and June 20^th, respectively.

Body weight (BW), testicular weight (TW) and plasma testosterone concentrations were recorded weekly throughout the study. Testicular weight was assessed by comparative palpation using an orchidometer; the same operator always performed this task (Oldham et al., 1978). Blood for determining the plasma testosterone concentration was obtained by jugular venipuncture, employing vacuum tubes containing heparin. This was performed once per week at 09:00 h over the entire experimental period. Plasma was obtained by centrifuging the collected blood at 3000 g for 30 min. It was then stored at -20ºC until analysis for testosterone using a commercial enzyme-linked immunoassay (ELISA) kit (Demeditec Diagnostics, Kiel-Wellsee, Germany). All plasma samples were analysed at the same time at the end of the experiment. The intra-assay coefficient of variation, estimated from plasma standards, was 8.9% for samples containing 0.5 ng/mL testosterone, 5.1% for samples with 2 ng/mL, and 7.3% for a samples containing 16 ng/mL.

Semen was collected weekly. On each occasion, the sexual behaviour of each buck was assessed by presenting it with an intact oestrus-induced doe, allowing 5 min for the male to ejaculate. Oestrus was induced in teaser does via a subcutaneous injection of 2 mg of oestradiol cypionate (Sigma-Aldrich Química, S.A., Spain) (Delgadillo et al., 1999; Zarazaga et al., 2009). The ejaculation latency, the percentage of bucks that ejaculated, and the percentage of active males (bucks that attempted to ejaculate at least twice, but did not achieve ejaculation within 5 min), were recorded. Animals were always tested in the same order and by the same handlers.

A total of 324 ejaculates were evaluated. The volume of ejaculated semen was recorded immediately after collection in a graduated collection vial. Overall motility was immediately assessed by transferring a drop of undiluted semen to a warm slide (35ºC), placing a cover slip on it, and observing it under a microscope at 40×. Results were recorded on an arbitrary scale of 0 to 5 (0 = no motility, 5 = 100% motility) (Baril et al., 1993).

Sperm concentration and kinematic motility variables were measured using a CASA system (ISAS, Proiser SL, Valencia, Spain). Sperm concentration was measured employing a Bürker chamber after diluting an aliquot of semen with a 0.05% formaldehyde saline solution (1:400) and observing at 400× magnification. The total number of spermatozoa per ejaculate was calculated from the ejaculate volume and sperm concentration. The kinematic motility variables recorded by the CASA system were: percentage of static, motile and progressive motile spermatozoa, curvilinear velocity (VCL, µm/s, i.e., the velocity of the actual trajectory of the sperm), straight-line velocity (VSL, µm/s, i.e., the velocity calculated using the straight-line distance between the beginning and end of the sperm track), average velocity (VAP, µm/s, i.e., the velocity over the smooth calculated path), and the linearity coefficient (LIN), straightness coefficient (STR) and wobble coefficient (WOB). Sperm were classified as being of medium velocity when their VCL was between 45 and 75 µm/s and rapid when their VCL was >75 µm/s. Sperm were deemed progressively motile when their trajectory was straight for at least 80% of the path taken, as suggested by the manufacturer of the CASA system.

Selected semen samples, i.e., those with a sperm concentration of >3500·10⁶ spermatozoa/mL, and an overall motility of ≥4 according to the above criteria, were frozen every fortnight. These criteria determined that a total of 128 ejaculates were processed. After checking that these criteria were met, the chosen samples were diluted in washing solution at 1:10 (v:v) (250 mM Tris, 28 mM glucose, 104 mM citric acid, 0.05% streptomycin, 500 UI penicillin/mL) at 37ºC, and then centrifuged once (700 g for 15 min) to eliminate the seminal plasma. The supernatant was then removed (now at room temperature [20ºC]) and the pellet diluted in a Tris-yolk extender (250 mM Tris, 28 mM glucose, 104 mM citric acid, 12% egg yolk, 0.05% streptomycin, 500 UI penicillin/mL and distilled water to 100 mL). After 5 min, a similar volume of a second extender (250 mM Tris, 28 mM glucose, 104 mM citric acid, 12% egg yolk, 8% glycerol, 0.05% streptomycin, 500 UI penicillin/mL, and distilled water to 100 mL) was added at room temperature (20ºC), resulting in a final sperm concentration of 800·10⁶ spermatozoa/mL, 6% egg yolk and 4% glycerol. Lecithin in the egg yolk was inactivated by subjecting the latter to 56ºC for 30 min before its use in the extenders. Both extenders were prepared in the laboratory using reagent-grade chemicals purchased from Panreac Química S.A. (Barcelona, Spain) and the Sigma Chemical Co. (St. Louis, MO, USA). The sperm was then chilled from room temperature to 5ºC in a cooler over a period of least 3 h, before being packed in 0.25 or 0.50 mL plastic straws. Finally, the straws were placed horizontally on a rack situated 4 cm above the surface of a liquid nitrogen bath for 15 min before being plunged into the same bath. The frozen straws were then stored in liquid nitrogen.

Sperm quality was evaluated after the 3 h chilling period and again after freezing-thawing; thawing was performed no later than one day after freezing. The frozen straws were thawed in a water bath at 37ºC for 30 s. Both the chilled and frozen-thawed sperm was transferred to a warm slide and the kinematic motility variables measured as described above.

The effect of season (spring, summer, autumn and winter) and week within the season (WWS, with 13 weeks per season) on BW, TW and plasma testosterone concentration was analysed using the repeated measures option of the General Linear Model procedure provided by the SPSS package for Windows (2008). In this model, the different seasons and WWS were taken as intra-subject factors. Significant differences in the influence of WWS were analysed using the Bonferroni test.

ANOVA including the male as fixed factor was used to examine the effect of season on all the studied variables in fresh, chilled and frozen-thawed sperm. Variables expressed as percentages (sperm motility, bucks that ejaculated, and percentage of active males), and the values for overall motility were arcsine-transformed before analysis. When differences between seasons were detected, they were examined using the Tukey t-test. ANOVA was also used to compare the effect of season on the motility variables of the rapid and medium velocity fresh, chilled and frozen-thawed sperm. Significance was set at p<0.05.

Repeated measures analysis showed season to have a significant effect on BW and plasma testosterone concentration (p<0.01). The BW was higher during spring than any other season, and plasma testosterone was higher during summer and autumn than during spring or winter (Table 1). Moreover, the interaction Season × WWS had a significant (p<0.01) effect on plasma testosterone, with concentrations increasing over the weeks of summer, and decreasing over the weeks of autumn. During spring and winter, testosterone concentrations remained at low levels, during summer testosterone concentrations increased, and during autumn they decreased (Fig. 1).

No differences were seen (Table 2) between seasons in terms of the percentage of bucks that ejaculated (85.6%), the percentage of bucks considered active (93.0%), ejaculation latency (47.0 ± 2.8 s), or ejaculate volume (0.91 ± 0.03 mL). However, differences were seen between seasons in terms of sperm concentration and the number of sperm per ejaculate, with values lower during winter and higher in summer (at least, p<0.01; Table 2). Finally, overall motility varied between seasons (p<0.01; Table 2), with the lowest values recorded during winter.

The VSL, LIN, STR and WOB values, and the percentages of motile, rapid and progressive spermatozoa, varied between season, with the lowest values recorded during winter (at least p<0.05; Fig. 2). For the rapid spermatozoa, only the values of VSL and LIN varied between seasons (Table 3), with the lowest values recorded in winter and the highest in summer. For the medium velocity spermatozoa, the values of all the kinematic motility variables varied between seasons, again with the lowest values recorded in winter and the highest in summer (Table 4).

Only the values of VCL, LIN, STR and WOB varied between seasons (at least p<0.05; Fig. 3), with the highest VCL and lowest LIN, STR and WOB values recorded in autumn. For the rapid spermatozoa, VSL, LIN, STR and WOB all varied between seasons, with the highest values recorded in spring (Table 3). For the medium velocity spermatozoa, the values of all the kinematic motility variables (except for VCL) varied between seasons, with the lowest recorded in autumn (Table 4).

The VCL and VAP results differed between seasons, with higher values recorded in winter than in summer. Very large differences were also observed in terms of the percentage of motile, rapid and progressive spermatozoa, with the best results recorded in winter (Fig. 4).

When examined separately, neither the rapid nor medium velocity spermatozoa differed between seasons in terms of any kinematic motility variable (Tables 3 and 4).