A commercial Norwegian dairy herd of 11 heifers and 78 cows was observed continuously for 21 days in October, during the housing season. The cow’s herd consisted of 78 cows in 1st, 2nd and over 2nd lactation, with a distribution of 43, 13, and 22 cows, respectively. A group of cows that could be expected to show estrus were selected for this study. Their milk was analyzed to measure progesterone concentration to confirm the occurrence of estrus. Cows selected for sampling were minimum 5 weeks postpartum. Exclusion criteria were reproductive disorders postpartum, insemination prior to the start of the study, pregnancy and moderate or severe lameness - a lameness score of over 2/5 in the week preceding the study’s start (21). To ensure that there would be cows in estrus during the study period, an estrus synchronization program consisting of two intra-muscular injections of 0.5 mg cloprostenol (Estrumate, MSD Animal Health) 11 days apart was commenced 21 days before the start of the study on cows that calved 7–10 weeks before study start. Progesterone sampling to confirm cyclicity was not performed after the treatment. The study group (n = 27) included 14, 2 and 11 cows in 1st, 2nd and over 2nd lactation, respectively.

The cows were housed in a free-stall barn with a concrete slatted floor, and 84 cubicles. The cows were milked twice daily, at 06:00–07:20 and 17:00–18:20, in a rotary parlor with 16 milking stalls. Forage was fed twice daily and consisted of brewer’s grains and grass silage ad libitum. Concentrates were fed, according to yield, by automated concentrate feeders placed in the free-stall area.

Samples for the determination of progesterone concentration in milk were taken by the farmer from cows in the selected study group, during every milking. Whole milk from included cows was sampled using a milk meter, ensuring samples representative for the entire milking event, and a bronopol tablet was thereafter added to the 20 mL sample. The samples were labeled immediately and kept at 4°C prior to freezing at −18°C on the day of collection. The samples were transported to the Norwegian School of Veterinary Science for analysis. Using enzyme immunoassay (22) and the second-antibody coating technique (23), the progesterone concentrations in the samples were determined. The samples were analyzed in duplicate, and the intra-assay coefficient of variation for progesterone concentration in whole milk was less than 10%. At progesterone concentrations in whole milk of 1.5 and 19.7 ng/mL, the inter-assay coefficients of variation were 9.2 and 5.3%, respectively. Using a 20 μL sample, the limit of sensitivity was less than 10%.

The 89 cows in the herd were marked individually prior to and during the study using white water paint on red and black hair coat and black paint on white hair coat every second or third day. The marking consisted of a number code unrelated to expected estrus and was the basis of the blinded study, in which the video observer had no knowledge of the estrous cycle stage of the individual study subjects. The behavior of all the cows in the herd was recorded. However, only events in which at least one cow from the selected study group (n = 18) was participating, is described in this study.

Five cameras (Axis 211, Axis Communications, Lund, Sweden) were placed to assure the whole area available to the cows was under surveillance. One picture was taken by each camera every second, stored as. jpg files, and saved on an internal hard drive in the camera. Photos were taken of every cow to be able to verify identification during video analysis. The program AXIS Camera Control (Axis Communications) was used to review the video files and record the activity of every cow according to the descriptors in section 2.5 (below) using a Microsoft Office Access Database (Microsoft Corp., Redmond, WA). The video was reviewed by two researchers individually, who then compared results. In the case of discrepancy both researchers looked at the video clip together and determined which behaviors were being displayed and which cows were involved. The final agreement represented the behavior and individual that was recorded in the final study dataset.

All mount and stand behaviors involve one initiating (I) and one receiving (R) cow. The definitions of mount and stand behaviors were modifications based on the definitions used by Sveberg et al. (24), and were as follows:

Each cow was defined to be in the luteal phase if the progesterone concentration in milk was 3 ng/mL or higher in samples from two consecutive milkings, or 5 ng/mL or higher in one sample (4, 22, 25, 26). The definitions of the different stages of the estrous cycle were as follows:

An illustration of the timeline of the different phases of the estrous cycle is presented in Figure 1.

Continuous variables were assessed for normality visually using histograms. The results of non-normally distributed variables are expressed as median values and interquartile range (IQR). Normally distributed variables are expressed as mean values ± standard deviation (SD). Data management and analyses was performed using Stata version 16.0 (StataCorp, College Station, Texas 77,845, USA).

Eighteen of the 27 selected cows were included in the study results. The reasons nine cows were excluded are as follows; one was slaughtered 7 days after the start of the study, two were excluded as purulent vaginal discharge was diagnosed in the study period, two cows’ samples were not analyzed due to missing information, three of the sampled cows were excluded because the study period started and/or ended in the middle of their estrus periods, and one cow was excluded due to sampling error.

The final study group (n = 18) included nine, two and seven cows in first, second and third or more lactation, respectively. The mean daily milk yield of the 18 cows included in the study was 24.4 ± 4.6 (±SD) kg milk. All the cows in the final study group were characterized as having cyclical ovarian activity based on milk progesterone concentration measurements, and all expressed MTE and STE.

From the first measured progesterone concentration < 3 ng/mL until the start of MTE, there was a median of 32 h, with a range from 20 to 87 h (IQR 26–47). The time from the end of MTE to the last measured progesterone concentration < 3 ng/mL was a median of 61 h, with a range from 25 to 94 (IQR 58–68). The median values of the durations of the estrus phases; MTE, STE, prestand and poststand, are presented in Table 1.

All the cows in the final study group expressed both MTE and STE. STBM and ‘all STBM’ accounted for 34 and 49% of all the mounts received during MTE, respectively. Four cows expressed STBM, and one cow expressed rSTBM as their first mounting activity during MTE (n = 13 for prestand). Two and three cows expressed STBM and rSTBM as their last mount activity during MTE, respectively (n = 13 for poststand). The total number of times a cow participated in mount activities during the estrus phases and the count per hour (cph) during these phases are presented in Table 2. In Table 3, the total numbers of initiated and received mounting and standing behaviors during the estrus phases are presented. Figure 2 presents the cph of initiated and received mounts and stand behaviors during prestand, STE and poststand.

During metestrus, nine of the 18 cows (50%) were mounted by another cow up to four times. These mounts resulted in a standing response once or twice for two of the 18 cows (11.1%). Six of the 18 cows (33.3%) in the study group mounted another cow once or twice during metestrus.

During proestrus, nine cows (50%) were mounted by another cow one to eight times in total. These mounts resulted in a standing response for six of the cows (33.3%), which stood to be mounted one to seven times during the same period.Of these STB events, 76.5% were unrestricted. Seven cows (38.9%) mounted another cow once or twice during proestrus.

In contrast to studies in other breeds all cows in this study expressed the primary sign of estrus, standing to be mounted (STBM), when in estrus (13, 14, 29). The number of times each cow expressed STBM during STE in this study was higher than previously reported in continuous studies for other breeds (14, 15, 30), but similar to the report for NR outdoors in Ireland (15). The high number of cows expressing STBM in this study indicates that onset of standing activity could be a more useful predictor of ovulation time for NR cows compared to HF cows where its usefulness has been questioned (29). However, STBM was observed during all four stages of the estrous cycle (estrus, metestrus, diestrus, proestrus) in this study, with more than 60% of the cows expressing the primary estrus sign during diestrus.

Standing to be mounted is traditionally considered the most reliable sign of a cow being in estrus (13, 18, 31) and has previously been reported rarely to happen when cows are not in estrus (29, 32). Traditionally the occurrence of STBM has been regarded as evidence of a cow being in estrus (13, 14). However, in this study such an interpretation would have led to many false-positive results, meaning that factors such as intensity of estrous signs should be considered. Inseminations during the luteal phase and during pregnancy have previously been reported (3, 4), indicating the occurrence of behaviors typically associated with estrus in pregnant cows and cows in diestrus. The reason for the observed high incidence of these behaviors during diestrus in the current study is unclear. Estrus expression is highly dependent on the concentration of circulating estradiol and the relationship between estradiol and progesterone (31, 33). One possible explanation could be that the cows were providing moderate milk yield (mean daily milk yield 24.4 ± 4.6 (±SD) kg) and therefore may have had follicles which produced more estradiol than those in higher yielding cows, or because of a lower steroid metabolism in relatively low yielding dairy cows (34). Further studies, preferably with transrectal ultrasound monitoring and estrogen and progesterone sampling throughout the study period, would be useful to better understand this finding.

Initiating mount behavior is considered to be an accurate sign of a cow being in or close to estrus (32), but has also been reported to occasionally occur during diestrus (13). In this study, cows expressed initiating mount behavior during all four phases of the estrous cycle, but with a considerably lower intensity during metestrus, diestrus and proestrus than during estrus. This suggests that the intensity of initiating mounts, in addition to expression of STBM, could be useful for correctly characterizing estrus in the NR cow. Esslemont and Bryant (35) have previously defined a cow as being in estrus if she mounted another cow six times or more in a 24 h period. More recently, in a study using continuous video monitoring to observe estrus behaviors in HF cows, the mean occurrence of initiated mounts was 6.6 (range 0–22) on the day of estrus (14). In this study, one cow initiated 97 mounts during diestrus. While the reasons for this are unknown it illustrates the potential importance of individual thresholds been set in precision farming estrus detection systems to optimize their sensitivity and specificity. Sensor technology has the possibility to measure every mount a cow makes. Therefore, it is important to investigate further the natural variation in mount behaviors throughout the estrous cycle under common conditions, to be able to adequately perform estrus detection with these devices.

The present study supports previous evidence that NR cows have long MTE and STE periods (15). The duration of MTE in this study (21 h) is considerably longer than mean duration of behavioral estrus reported for HF cows in previous studies, of between 11 and 14 h (13, 14, 29). While the duration of STE (14 h) in this study population is similar to the durations reported for HF cows in the seventies, eighties and early nineties (13, 24, 29), but longer than the more recent results (36). Milk yields are negatively associated with the duration of standing estrus (9) and so the lower production of NR cattle could partially explain the reduction in the duration of estrus in HF compared to NR cows. However, in a study comparing HF and NR differences in estrus duration were present, despite milk yield and environmental conditions been similar (15).

The reason for this difference has not been investigated by this study. However, the similarity of the current NR’s STE to those of historic studies in HF is intriguing and requires further investigation. Our hypothesis is that including reproductive traits in the breeding program of the NR since the 1970s (37) combined with the low use of reproductive synchronization programs in Norway (38) has meant that the duration of estrus has been preserved, as it has been critical to breeding success in Norway. Further work is required to test this hypothesis.

The high cph found in this study suggests a high intensity of mount behaviors in NR during estrus, despite being housed on concrete floors, which have been negatively associated with such behaviors (6, 19). However, because hormonal estrus synchronization was performed prior to the study, it is reasonable to assume that there might have been a higher number of cows in estrus at the same time, than in a normal situation in a commercial farm. This could have skewed the results toward a higher intensity of estrus expression due to the formation of sexually active groups. Despite this, the intensity of estrus signs, measured as median counts per hour (cph) of mount events during MTE, in this study were similar to the means previously reported for NR (15). A long duration and intensity of estrus could make estrus easier to detect during traditional observing periods but is also important to consider when adapting estrus detection technologies for use in NR herds. However, cph is an average measure and therefore does not reflect any periods of higher or lower intensity during the defined period. It is therefore possible that periods of intense mounting and standing activity could be missed by farmers depending on visual estrus detection, particularly if they do not spend sufficient time.

The cows in this study were kept under typical conditions for dairy cows in Norway and were minimally disturbed during data collection, as estrus detection was based on video surveillance. A strength of this study is that the housing environment and management of the herd was typical for NR herds in Norway. This contrasts with the study previously performed comparing estrus in NR and HF cows in Ireland where the cows were kept outdoors in a restricted area and in an experimental setting (15). Further, cows across many parities were included in this study, although no heifers were included.

Undoubtedly the study was limited by the low number of animals remaining in the study group after exclusions were made. The small numbers prevented meaningful statistical analysis from being performed, making the study descriptive in nature. Further, only healthy cows were included in the study. A number of diseases are known to negatively influence display of estrus, e.g., lameness (39), so by selecting a healthy population the study does not accurately represent a commercial dairy herd, with inherent endemic disease. Unfortunately, the effect of number of cows in estrus on mount and stand activity and on the participation of diestrous cows in sexually active groups (40) could not be evaluated in the current study, due to the high number of cows with unknown cycle status in the herd. It is possible that by synchronizing the cows before the study start that an ‘artificial’ sexually active group was created. Further studies, including all cows in a herd, are needed to better understand these behaviors.

Despite these limitations the findings of the current study largely support the previous findings for NR by Sveberg et al. (15) and we believe the results of this study are applicable to commercial NR dairy cows.