See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325470954

Association of production diseases with motor activity-sensing devices and

milk progesterone concentrations in dairy cows

Article  in  Theriogenology · May 2018

DOI: 10.1016/j.theriogenology.2018.05.038

CITATIONS READS

0 32

9 authors, including:

Jess A. Williams Theodoros Ntallaris

Liverpool School of Tropical Medicine Swedish University of Agricultural Sciences
5 PUBLICATIONS   69 CITATIONS    14 PUBLICATIONS   18 CITATIONS   

SEE PROFILE SEE PROFILE

Robert Frank Smith Patrice Humblot

University of Liverpool Swedish University of Agricultural Sciences
102 PUBLICATIONS   3,092 CITATIONS    262 PUBLICATIONS   3,361 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Propylene Glycol in donor cows View project

DetOestrus View project

All content following this page was uploaded by Jess A. Williams on 25 July 2019.

The user has requested enhancement of the downloaded file.

 Theriogenology 118 (2018) 57e62

Contents lists available at ScienceDirect

Theriogenology
journal homepage: www.theriojournal.com

Association of production diseases with motor activity-sensing

devices and milk progesterone concentrations in dairy cows
J. Williams a, T. Ntallaris b, J.E. Routly a, D.N. Jones a, J. Cameron a, A. Holman-Coates a,
R.F. Smith a, P. Humblot b, H. Dobson a, *
a
Leahurst Campus, School of Veterinary Science, University of Liverpool, Neston, CH64 7TE UK
b
Division of Reproduction, Department of Clinical Sciences, Faculty of Veterinary Medicine and Agricultural Sciences, SLU, 750-07 Uppsala, Sweden

a r t i c l e i n f o a b s t r a c t

Article history: We have previously established that the efficiency of identifying oestrus with activity-sensing devices
Received 13 January 2018 can be compromised by common production diseases; the present study was undertaken to determine
Received in revised form how these diseases may affect device readings. A total of 67 Holstein-Friesian cows, >20 days post-
24 May 2018
partum, were equipped with activity-sensing neck collars and pedometers, and simultaneous milk
Accepted 29 May 2018
Available online 31 May 2018
progesterone profiles were also monitored twice a week. The influences of common production stressors
on maximum activity and progesterone values were analysed. Approximately 30% potential oestrus
events (low progesterone value between two high values) remained unrecognised by both activity
Keywords:
Oestrus
methods, and progesterone values in these animals were higher on the potential day of oestrus when
Lameness both activity methods did not detect an event (0.043 ± 0.004 versus 0.029 ± 0.004 ng/mL; P ¼ 0.03). Data
Body condition score from a subset of 45 cows (two events each) were subjected to mixed models and multiple regression
modelling to investigate associations with production diseases. Cow motor activity was lower in lame
cows. Maximum progesterone concentrations prior to oestrus increased as time postpartum and body
condition score (BCS) increased. There were also fewer days of low progesterone prior to oestrus asso-
ciated with increases in BCS and maximum progesterone concentrations prior to oestrus. In conclusion,
lameness was associated with lower activity values, but this suppression was insufficient to account for
lowered oestrus detection efficiency of either device. However, associations were identified between
production diseases and progesterone profiles.
© 2018 Elsevier Inc. All rights reserved.

1. Introduction farm staff on dairy farms, the popularity of activity-sensing devices

There has been a decline in oestrus behaviour intensity dis-
played by dairy cattle world-wide over the past 50 years leading to
problems with accurate oestrus detection in many herds [1].
Oestrus normally lasts 15e18 h, although some cattle show signs
for as little as 6 h, making identification a challenge for busy farm
staff [2].
From a management point of view, being able to efficiently and
accurately detect oestrus is still the most important factor in
reducing calving intervals [1,3]. Accuracy of oestrus detection
should be examined as part of evaluation of herds with low preg-
nancy rates, especially as a visual heat detection rate of only 38%
was recorded across 4550 dairy herds [4]. These days with fewer
is increasing. Activity-sensing neck-collars or pedometers both
engage continuous radio-transmitted monitoring of cow activity,
and have been reviewed across many studies with an average 87%
accuracy when compared to visual observations [5,6].
A prior comparison of various simultaneous methods of oestrus
detection indicated that the efficiency of identifying oestrus is
compromised by common production diseases such as lameness,
metabolic disorders as revealed by low body condition score (BCS),
and high daily milk yield [7]. The present study aimed to establish if
production stressors affected specific components of two types of
cow activity devices (neck-collars or pedometers). Associations
with parity and milk constituents (somatic cell counts (SSC), % fat,
and/or % protein, fat: protein ratio) were also examined.

* Corresponding author.
E-mail address: hdobson@liv.ac.uk (H. Dobson).

https://doi.org/10.1016/j.theriogenology.2018.05.038
0093-691X/© 2018 Elsevier Inc. All rights reserved.
58 J. Williams et al. / Theriogenology 118 (2018) 57e62
2. Materials and methods
2.1. Animals
This study was performed under a UK Home Office license for
work on living animals and with the approval of the University of
Liverpool Ethical Review Process. HolsteineFriesian cows in their
first to eleventh lactation entered the six-month study (March-
eAugust) of a 200-cow herd that calved throughout the year on one
free-housed commercial UK dairy farm. Mean milk sales per cow
were around 11,000 L annually, with average peak yields of 54 L per
day. At any one time, a group of approximately 80 cows in the
period four months after calving were housed in a cubicle shed
with concrete passage-ways. Milking of all cows took place three
times a day starting at 05:30 h, 14.00 h and 20:30 h. All animals had
continual access to a total mixed ration that was pushed up to a
feed-fence five times a day. All cows were monitored after entering
the study until confirmed pregnant after ultrasound examination
by the farm's veterinarian five weeks after artificial insemination or
until the end of the study period. Therefore, many animals had
more than one potential oestrus event studied.
Every two weeks, the same member of the research team
assessed each cow's body condition score (BCS) using a 1e5 system
incorporating 0.5 scores [8], and lameness using a standardised 1e5
system [9]. For statistical analysis, individuals were categorised as
not lame (lameness score <1.5) or lame (lameness score >1.6;
n ¼ 24 cows). Once a month, individual cow 24 h milk yields
(subsequently converted to ECM: energy corrected milk (4% fat and
3.3% protein calculated as 24 h milk production  0.383  fat
% þ milk production  0.242  protein % þ 0.7832  milk pro-
duction/3.1138), somatic cell counts (SCC), % fat and % protein were
recorded by Cattle Information Services (CIS; Rickmansworth, UK).
The monthly value immediately prior to a potential oestrus event
(NEAR_) was used for analysis. Treatments for lameness and
mastitis followed normal farm practice: regular or remedial foot-
trimming, and antibiotic treatment for udder infections.
2.2. Activity-sensing devices
On entry into the study, all cows were fitted with an activity-
sensing neck-collar (Heatime™; SCR Engineers, Netanya, Israel;
distributed by FabDec, Ellesmere, UK) with data being downloaded
to a dedicated program (Dataflow, SCR Engineers) in 2-h blocks at
the milking parlour exit. As designated by the equipment provider,
a value of greater than 4.7 standard deviations above the mean
activity of the previous eight two-hourly blocks was used to indi-
cate oestrus - this algorithm precluded assessment of baseline
values. Also at recruitment, a leg pedometer (Afitag; SAE Afikim,
Israel; distributed by Fullwood, Ellesmere, UK) was placed on all
cows as an additional measure of cow activity. The pedometer data
were also collated remotely by a computer in 8-h blocks at milking
using a dedicated program (Crystal; Fusion Electronics, Fullwood).
Table 1
Efficiencies of detection for neck-collars, pedometers or combined device data, with accompanying milk progesterone concentrations on day of potential oestrus
(mean ± SEM).
Device Total potential Events Efficiency Milk progesterone (ng/mL) for
events (n) recorded (n) (%) recorded event
Neck- 137 92 67 0.035 ± 0.004
collar
Pedometer 151 106 70 0.032 ± 0.003
Combined 164 72 45 0.029 ± 0.004
The algorithm used to identify an oestrus was based on an increase
in steps per hour greater than 70% of the average for the previous 10
days.
2.3. Hormone analyses
Milk samples for progesterone analysis were collected each
Tuesday and Friday throughout the study. All samples were taken
immediately before milking and promptly stored at 20  C without
preservative. Progesterone concentration was analysed as pregnane
metabolites in 50 ml whole milk samples using an established EIA
assay [10]. For the present study, the minimum detectable amount
was 0.02 ng/mL, and the intra- and inter-assay coefficients of
variation were 9.7 and 15.8%, respectively. When values were less
than 0.2 ng/mL, luteal tissue was considered to be absent, whereas
values greater than 0.3 ng/mL indicated the presence of luteal tis-
sue [10]. Progesterone profiles were produced for each study ani-
mal and used to indicate potential oestrus periods to be identified
by neck-collars or pedometers. The term “potential oestrus event”
used in data analysis was defined as each time milk progesterone
concentration was below 0.2 ng/mL for 2e6 days, preceded and
followed by an increase above 0.3 ng/mL for a minimum of 7e10
days.
2.4. Data analyses
Sixty-seven cows were studied providing 164 potential oestrus
events for consideration. Data from four cows were excluded
because the progesterone profiles were very irregular. Data were
also excluded from 34 potential events due to obvious false posi-
tives within the neck collar or pedometer data (which mainly
occurred during the luteal phase). The neck collar system mal-
functioned in 31 (18%) and the pedometers malfunctioned in 17
(10%). These malfunctions were due to mixtures of human error,
e.g., power outage, malfunction of the equipment, failure to
correctly enter information on the computers. Statistical analysis
allowed for these missing data.
Milk progesterone concentrations during all potential oestrus
events identified by either neck collars or pedometers were initially
compared by Students t-test. Due to possible bias and problems
introduced either by over-representation or under-representation
of some cows while running Mixed Models, subsequent analyses
were performed using the first two potential oestrus events in a
total of 45 cows which contributed to the final balanced data-set.
This comprised 10 primiparous animals, 14 cows that had had 2
calvings, 11 had 3 calvings, 4 had 4 calvings and six cows had had 5
or more calvings).
All statistical analyses were performed with SAS software
version 9.3, using the MIXED procedure for linear Mixed Models,
unless stated differently. A repeated effect of time (day postpartum)
within animals was tested. The correlations between test days were
accounted for by specifying a correlation structure (spatial power)
Milk progesterone (ng/mL) for P (between progesterone
missed event concentrations)
0.051 ± 0.007 0.090
0.042 ± 0.007 0.059
0.043 ± 0.004 0.028
 J. Williams et al. / Theriogenology 118 (2018) 57e62 59

among residuals to account for differing intervals between
samplings.
Stepwise regression was used with a backward elimination
approach to fit the final regression model. In each step a factor was
kept or excluded from the set of explanatory ones based on pre-
specified exclusion criterion of non-significant (P > 0.20) effects.
The residuals from the response variables generated from the
model were tested for normal distribution. Progesterone values at
oestrus, and the number of prior days with low progesterone
values, deviated from a normal distribution and were log-
transformed. To facilitate interpretation of the data for these pa-
rameters, non-transformed values are presented.
Values for each milk variable were taken as those measured
nearest to each potential oestrus event; other dependent variables
were: maximum neck-collar value, MAX_HT; maximum pedometer
value, MAX_PEDO; baseline pedometer value, BASE_PED; proges-
terone value at oestrus, PROG_OE; the number of days prior to
oestrus with low progesterone values, DAYS_PRI; maximum pro-
undetected event (overall mean ± SEM: 3.82 ± 0.43 days; P ¼ 0.23).
However, progesterone concentrations on the day of potential oestrus
events that were not identified by either neck collars or pedometers
tended to be greater than for events registered by each device
(P ¼ 0.090 and 0.059, respectively; Table 1). When data were com-
bined to examine only those events identified by the two devices
simultaneously, progesterone concentrations were significantly
greater for events undetected by both devices (P ¼ 0.028; Table 1).
3.2. Associations between cow activity, progesterone measurements
and production diseases
Possible associations between cow activity, progesterone
500
A

Maximum pedometer value

gesterone concentrations in the luteal phase prior to a potential 400
P=0.023
oestrus, MAX_PRIO.

(arbitray units)
The initial model used was: 300

Yijklmnoqr ¼ m þ ai þ bj þ ck þ dl þ fm þ gn þ ho þ Iq þ eijklmnoqr 200

Where Yijklmnoqr is the dependent variable (MAX_HT; MAX_PEDO, 100

BASE_PED, PROG_OE, DAYS_PRI, MAX_PRIO), m is the overall mean,
ai is the fixed effect of postpartum oestrus number i (i ¼ 3 classes; 0
PP_OE), bj is the fixed effect of oestrus on day postpartum j No Yes
(j ¼ 27e130; DAY_PP), ck is the fixed effect of energy corrected milk lame
production near oestrus k (k ¼ 28.3e57.9; ECM), dl is the fixed ef-
fect of fat: protein ratio near oestrus (l ¼ 0.6e1.9; NEAR_FP), fm is
the fixed effect of SCC near oestrus m (m ¼ 9.0e2264; NEAR_SCC), 500
gn is the fixed effect of BCS near oestrus n (n ¼ 1.8e3.0; BCS), ho is B
the fixed effect of lameness o (o ¼ 2 classes; lame yes/no ¼ lame
Baseline pedometer value

400
score <1.5 or >1.6), Iq is the fixed effect of parity q (q ¼ 2 classes; 2
(arbitrary values)

or >2), and eijklmnoqr is the residual error term. Scheffe's adjustment

300
was used for multiple-post ANOVA comparisons. Results are pre-
sented as least square means ± standard error of means
(LSM ± SEM) unless otherwise stated. 200
P=0.016
Interrelationships between variables were further analysed
from general linear models (Proc GLM), and MANOVA was used to 100
obtain partial correlations for the studied variables. This last model
included the fixed effects of parity from initial raw data (9 classes). 0
There was a significant effect of parity on most variables, thus all No Yes
relationships/correlations in the model were accordingly adjusted lame
using the MANOVA statement in GLM procedure of SAS. Differences
were considered significant when P < 0.05, with a tendency being
between 0.05 and 0.15 for MANOVA correlations.
250
3. Results C
200
Maximum neck-collar value

Excluding spasmodic equipment malfunctions as explained P=0.141

(arbitrary units)

above in Data Analysis, neck-collars identified 67% potential oestrus 150

events; similarly, pedometers detected 70% events (Table 1). Thus,
approximately 30% potential events remained undetected by either 100
activity device.
50
3.1. Relationship between milk progesterone concentrations and
increased cow activity 0
NO Yes
In the full data-set (a total of 164 potential events), maximum lame
progesterone concentrations prior to an identified potential oestrus
event were not different (P ¼ 0.89) to those prior to undetected events Fig. 1. Effect of lameness (no/yes) on maximum and baseline pedometer values (A and
for both devices (overall mean ± SEM: 1.11 ± 0.07 ng/mL); nor were B, respectively) and maximum neck collar values (C). Results are presented as least
the number of days with low progesterone prior to an identified or square means ± SEM.
60 J. Williams et al. / Theriogenology 118 (2018) 57e62

measurements and production diseases (as defined previously) these animals tended to have more days with low progesterone
were then examined in a balanced data-set that included the two concentrations prior to oestrus (Fig. 2F).
first postpartum potential oestrus events from 45 cows (i.e., 90 Relationships between all variables were further considered by
events). examining partial correlations (MANOVA model), and the results
A total of 24 cows had a lameness score >1.6. From MIXED are summarized in Table 2. There were few relationships between
model analysis, lameness was associated with lower maximum and production disease variables; fat: protein ratio (NEAR_FP) was
baseline pedometer values (P < 0.05), and a tendency for lower positively associated with Energy Corrected Milk (ECM), and BCS
maximum neck-collar values; Fig. 1. Lameness was also associated was positively associated with oestrus number (PP_OE; P ¼ 0.01).
with lower progesterone concentrations at oestrus (P < 0.01; Maximum progesterone concentrations prior to oestrus were
Fig. 2A) and a similar tendency with maximum concentrations of positively associated with progesterone concentrations at oestrus,
progesterone prior to oestrus (Fig. 2B). Oestrus number (PP_OE) but negatively associated with number of days with low concen-
was positively associated with maximum progesterone concen- trations prior to oestrus (P < 0.03).
trations prior to oestrus but negatively associated with the number Maximum pedometer values were positively associated with
of days for which progesterone was low before oestrus (for both baseline pedometer values (P < 0.0001) and with maximum neck-
P ¼ 0.01; Fig. 2C and D). Finally, progesterone concentrations at collar values (P < 0.0001). Overall, only tendencies were found for
oestrus were higher in cows of parity >2 (P ¼ 0.0015; Fig. 2E), and relationships between activity variables and progesterone

0.07
A 1.6 B

Maximum prior progesterone ng/ml

0.06 1.4
Progesterone at oestrus ng/ml

0.05 1.2
p=0.008
1
0.04 p=0.13
0.8
0.03
0.6
0.02
0.4
0.01 0.2

0 0
lame, no lame, yes lame, no lame, yes

1.6 12
C D
Maximum prior progesterone ng/ml

1.4
Prior days with low progesterone

10
1.2 p=0.011
8
1
p<0.01
6
0.8
0.6 4
0.4
2
0.2
0
0 1 2 >=3
1 and 2 3
PP_OE PP_OE

0.07 12
E F
Prior days with low pogesterone
Progesterone at oestrus ng/ml

0.06 10
0.05 p=0.0015 p=0.13
8
0.04
6
0.03
4
0.02
2
0.01
0
0 parity > 2
parity > 2

Fig. 2. Effects of categorical variables on progesterone concentrations at potential oestrus (A and E), maximum concentrations of progesterone prior to potential oestrus (PP_OE; B
and C), and number of days with low progesterone concentrations prior to potential oestrus (D and F). Results are presented as least square means ± SEM.
 J. Williams et al. / Theriogenology 118 (2018) 57e62 61

Table 2
Partial Correlation Coefficients from the Error SSCP Matrix (upper row) with corresponding P value presented below each variable. The Partial correlation coefficients were
adjusted for an overall parity effect (P < 0.00001).

r/p-value DAY_PP ECM NEAR_FP NEAR_SCC BCS lame PROG_OE DAYS_PRI MAX_PRIO MAX_HT MAX_PEDO BASE_PED

PP_OE 0.47 0.02 0.04 0.02 0.34 0.24 0.002 0.23 0.32 0.15 0.15 0.01
<0.001 ns ns ns 0.01 0.07 ns 0.09 0.02 ns ns ns
DAY_PP 1 0.02 0.10 0.02 0.13 0.19 0.10 0.14 0.01 0.27 0.05 0.05
ns ns ns ns ns ns ns ns 0.04 ns ns
ECM 1 0.36 0.09 0.06 0.17 0.02 0.13 0.03 0.09 0.05 0.20
0.01 ns ns ns ns ns ns ns ns 0.14
NEAR_FP 1 0.01 0.23 0.10 0.20 0.14 0.04 0.03 0.19 0.34
ns 0.08 ns 0.14 ns ns ns 0.15 0.01
NEAR_SCC 1 0.21 0.07 0.13 0.07 0.06 0.06 0.30 0.26
0.13 ns ns ns ns ns 0.03 0.05
BCS 1 0.18 0.07 0.28 0.26 0.01 0.005 0.10
ns ns 0.04 0.05 ns ns ns
Lame score 1 0.29 0.23 0.16 0.01 0.00 0.13
0.03 0.09 ns ns ns ns
PROG_OE 1 0.06 0.34 0.21 0.24 0.18
ns 0.01 0.13 0.08 ns
DAYS_PRI 1 0.28 0.16 0.05 0.05
0.03 ns ns ns
MAX_PRIO 1 0.07 0.06 0.06
ns ns ns
MAX_HT 1 0.57 0.22
<.0001 0.10
MAX_PEDO 1 0.58
<.0001

ns: P value > 0.05; P values 0.05 to 0.15 indicate a tendency.

measurements. Few relationships were found between production
diseases and activity measurements.
Relationships between cow status, production diseases and
progesterone measurements were associated with BCS: higher BCS
being associated with fewer days of low progesterone before oes-
trus, and higher progesterone concentrations before oestrus (both
P  0.05). Furthermore, increased lameness was associated with
lower progesterone concentrations at the time of oestrus (P < 0.05),
and there tended to be more days with low progesterone prior to
oestrus (P < 0.10).
4. Discussion
The present study confirms that a large proportion (~30%) of
periods with low milk progesterone values in postpartum cows are
not associated with increased motor activity, although progester-
one profiles indicated regular ovarian cyclicity [7]. That is, some
cows did not display signs of oestrus possibly as a behavioural
trade-off while exposed to adverse conditions. Nevertheless, there
was no difference in efficiency between neck-collars and pedom-
eters to detect periods of low progesterone. The maximum activity
values recorded in the present study were considerably greater
than the threshold values for both devices and so do not fully
explain why many potential oestrus periods were not identified.
However, cows that were not detected did have marginally higher
milk progesterone concentrations at a potential oestrus than those
detected with increased motor activity. Moreover, following ad-
justments for the effects of all concurrent variables, lameness was
associated with lower activity values. In addition, there were sig-
nificant correlations between three indices of progesterone con-
centrations (lower values at oestrus, lower maximum values prior
to oestrus, and number of days with low values prior to oestrus)
and the following: day postpartum (and postpartum oestrus
number), parity, lameness and BCS. In this high-yielding herd, no
associations with cow activity or progesterone variables were
identified with ECM, although ECM was positively associated with
fat: protein ratios.
The association of lameness with lower activities of both
detection devices coincides with visual observations that lame
cows walk less, lie down more, and have reduced oestrus intensity
once ovarian cyclicity has resumed after calving [10]. These re-
ductions may be due to the physical pain associated with lameness,
however, other (hormonal) influences are involved. Although
ovarian follicular emergence, selection and physical growth are
unaffected, lame cows have low LH pulse frequencies in the late
follicular phase, consequently low simultaneous plasma oestradiol
concentrations, display less intense oestrus behaviour and have a
reduced incidence of LH surges and ovulations [11]. There were also
markedly lower progesterone concentrations during the luteal
phase when severely lame cows were compared to non-lame ani-
mals. In the present study, lameness was considered in two cate-
gories (lame/not lame) but there was still a tendency for lameness
to be associated with lower maximum progesterone values prior to
oestrus concurring with the hypothesis that lameness disturbs
normal ovarian function [10].
We have already established that lame cows have a lower bite
rate at pasture and are more likely to have a low BCS [12]. The
negative correlation of BCS with the number of days with low
progesterone prior to oestrus (i.e., as BCS increases the follicular
phase becomes shorter) agrees with the hypothesis that negative
energy balance (low BCS) affects gonadotrophin pulsatility and
ovarian steroid synthesis [13,14]. Similarly, a healthy follicle will
produce a good corpus luteum (as BCS increased, so did maximum
prior progesterone values). However, some cows cannot cope with
negative energy balance after calving, and as suggested many years
ago, any additional stressor (such as lameness) that disturbs this
knife-edge balance will ultimately lead to impaired reproductive
function including poor oestrus expression [15,16].
Milk fat and milk protein values were excluded from the final
statistical model used in the present study as having non-
significant influence. However, fat: protein ratios tended to be
positively correlated with BCS and progesterone values at oestrus.
Previously, cows with a first test-day fat: protein ratio of >1.5 have
been associated with higher risks for clinical disease, including
ketosis which has a negative effect on oestrus detection [17e19].
There is now the potential to harness the data from newly available
62 J. Williams et al. / Theriogenology 118 (2018) 57e62
in-line commercial devices that greatly increase the frequency of
milk analyses, and allow each cow to serve as its own control over
time. This will avoid the large variation in milk constituents both
within and between individuals [20].
An increase in maximum progesterone values, and fewer days
with low progesterone, prior to oestrus occurred as postpartum
oestrus number increased. These follow expected patterns as
ovarian (and hypothalamic-pituitary) function resume maturity
after calving [15,21]. Similarly, the number of days prior to oestrus
with low progesterone and progesterone concentrations at oestrus,
both increased as parity increased. Progesterone concentrations
prior to and on the day of oestrus are important factors associated
with behavioural manifestations of oestrus, and distorted values in
lame cows are associated with abnormal oestrus expression
[10,22].
In conclusion, production stressors are associated with differ-
ences in activity-sensing maximum values but, realistically, the
differences were insufficient to explain the low detection efficiency
of either device studied. This confirms conclusions regarding
higher milk yields associated with maximum values from similar
pedometers [23]. However, the present study indicates that asso-
ciations between production diseases and progesterone profiles
might be of considerable importance.
Conflicts of interest
None of the authors of this manuscript has any financial or
personal relationship with other people or organizations that could
inappropriately influence or bias the content of the manuscript.
Contributions
All authors designed the study, acquired and handled the data,
and approved the final manuscript.
Acknowledgements
This work was carried out by Jessica Williams in partial fulfil-
ment of the requirements for the Masters in Animal Reproduction,
University of Liverpool, UK; and Theodoros Ntallaris in partial
fulfilment of the requirements for a PhD, SLU, Sweden. All the au-
thors are grateful to the farm staff for their help during all hours of
the day and night. This research did not receive any specific grant
from funding agencies in the public, commercial, or not-for-profit
sectors.
References
[1] Dobson H, Walker SL, Morris MJ, Routly JE, Smith RF. Why is it getting more
difficult to successfully artificially inseminate dairy cows? Animal 2008;2:
View publication stats
1104e11.
[2] Dobson H, Smith R, Royal MD, Knight CH, Sheldon IM. The high-producing
dairy cow and its reproductive performance. Reprod Domest Anim 2007;42:
17e23.
[3] Britt JH, Scott RG, Armstrong JD, Whitacre MD. Determinants of oestrus
behaviour in lactating dairy cows. J Dairy Sci 1986;69:2195e202.
[4] Heersche J, Nebel RL. The accuracy of enzyme-linked immunosorbent assay
and latex aggulination progesterone test for the validation of estrus and early
pregnancy diagnosis in dairy cattle. Theriogenology 1994;39:1121.
[5] Cavalieri J, Flinker LR, Anderson GA, Macmillan KL. Characteristics of oestrus
measured using visual observation and radiotelemetry. Anim Reprod Sci
2003;76:1e12.
[6] Kaim M, Ertracht S, Lavon Y, Lehrer AR. Determination of insemination time in
PG-treated dairy heifers by activity monitoring and/or visual oestrous detec-
tion. J Dairy Sci 2008;86:2012e21.
[7] Holman AJ, Thompson J, Routly JE, Cameron J, Jones DN, Grove-White D,
Smith RF, Dobson H. Comparison of oestrus detection methods in dairy cattle.
Vet Rec 2011;169:47e54.
[8] Edmonson AJ, Lean IJ, Weaver LD, Farver T, Webster G. A body condition
scoring chart for Holstein dairy cows. J Dairy Sci 1989;72:68e78.
[9] Sprecher DJ, Hostetler DE, Kaneene JB. A lameness scoring system that uses
posture and gait to predict dairy cattle reproductive performance. Ther-
iogenology 1997;47:1179e91.
[10] Walker SL, Smith RF, Jones DN, Routly JE, Dobson H. Chronic stress, hormone
profiles and estrus intensity in dairy cattle. Horm Behav 2008a;53:493e501.
[11] Morris MJ, Kaneko K, Walker SL, Jones DN, Routly JE, Smith RF, Dobson H.
Influence of lameness on follicular growth, ovulation, reproductive hormone
concentrations and oestrus behaviour in dairy cows. Theriogenology 2011;76:
658e68.
[12] Walker SL, Smith RF, Routly JE, Jones DN, Morris MJ, Dobson H. Lameness,
activity time-budgets and estrus expression in dairy cattle. J Dairy Sci
2008b;91:4552e9.
[13] Butler WR. Nutritional interactions with reproductive performance in dairy
cattle. Anim Reprod Sci 2000;60e61:449e57.
[14] Mellouk N, Rame C, Touze
JL, Briant E, Ma L, Guillaume D, Lomet D, Caraty A,
Ntallaris T, Humblot P, Dupont J. Involvement of plasma adipokines in
metabolic and reproductive parameters in Holstein dairy cows fed with diets
with differing energy levels. J Dairy Sci 2007;100:8518e33.
[15] Dobson H, Smith RF. What is stress, and how does it affect reproduction?
Anim Reprod Sci 2000;60e61:743e52.
[16] Peake KA, Biggs AM, Argo CM, Smith RF, Christley RM, Routly JE, Dobson H.
Effects of lameness, sub-clinical mastitis and loss of body condition on the
reproductive performance of dairy cows. Vet Rec 2011;168:301e9.
[17] Heuer CY, Schukken H, Dobbelaar P. Postpartum body condition score and
results from the first test day milk as predictors of disease, fertility, yield and
culling in commercial dairy herds. J Dairy Sci 1999;82:295e304.
[18] Gillund P, Reksen O, Gro €hn YT, Karlberg K. Body condition related to ketosis
and reproductive performance in Norwegian dairy cows. J Dairy Sci 2001;84:
1390e6.
[19] Rutherford AJ, Oikonomou G, Smith RF. The effect of subclinical ketosis on
activity at estrus and reproductive performance in dairy cattle. J Dairy Sci
2016;99:1e8.
[20] Tsenkova R, Atanassova S, Itoh K, Ozaki Y, Toyoda K. Near infrared spectros-
copy for biomonitoring: cow milk composition measurement in a spectral
region from 1,100 to 2,400 nanometers. J Anim Sci 2000;78:515e22.
[21] Royal MD, Darwash AO, Flint APF, Webb R, Woolliams JE, Lamming GE.
Declining fertility in dairy cattle: changes in traditional and endocrine pa-
rameters of fertility. Anim Sci 2000;70:487e501.
[22] Mondal MA, Rajkhowa AC, Prakash BS. Relationship of plasma oestradiol 17 b,
total oestrogen, and progesterone to oestrus behaviour in mithin (Bos fron-
talis) cows. Horm Behav 2006;49:626e33.
[23] Lopez-Gatius F, Santolaria P, Mundet I, Yaniz JL. Walking activity at estrus and
subsequent fertility in dairy cows. Theriogenology 2005;63:1419e29.