Variations in the vulvar temperature of sows during [613113]

Variations in the vulvar temperature of sows during
proestrus and estrus as determined by infraredthermography and its relation to ovulation
Vasco G. Simõesa,1, Faouzi Lyazrhia, Nicole Picard-Hagena,b,
Véronique Gayrarda,b, Guy-Pierre Martineaua, Agnès Waret-Szkutaa,c,*
aInstitut National Polytechnique-Ecole Nationale Vétérinaire de Toulouse (INP-ENVT), Toulouse, France
bUnité Mixte de Recherche Institut National de la Recherche Agronomique-Ecole Nationale Vétérinaire de Toulouse (INRA-ENVT)
1331, Toxalim, Research Center in Food Toxicology, Toulouse, France
cUnité Mixte de Recherche Institut National de la Recherche Agronomique-Ecole Nationale Vétérinaire de Toulouse (INRA-ENVT)
1225, Interactions Hôtes –Agents Pathogènes (IHAP), Toulouse, France
article info
Article history:
Received 5 January 2014Received in revised form 8 July 2014Accepted 8 July 2014
Keywords:
SowSwine
Estrus
OvulationThermographyReproductionabstract
The prediction of ovulation time is one of the most important and yet dif ficult processes in
pig production, and it has a considerable impact on the fertility of the herd and litter size.
The objective of this study was to assess the vulvar skin temperature of sows duringproestrus and estrus using infrared thermography and to establish a possible relationship
between the variations in vulvar temperature and ovulation. The experimental group
comprised 36 crossbred Large White /C2Landrace females, of which 6 were gilts and 30
were multiparous sows. Estrus was detected twice daily and the temperature was obtainedevery 6 hours from the vulvar area and from two control points in the gluteal area (Gluteal
skin temperature [GST]). A third variable, vulvar –gluteal temperature (VGT) was obtained
from the difference between the vulvar skin temperature and the GST values. The animalswere divided into two subgroups: group A consisting of 11 animals with estrus detected at
6:00 AM, Day 4 postweaning, and group B comprising seven animals with estrus detected
at 6:00 AM, Day 5 post-weaning. Both groups showed a similar trend in the VGT. The VGTincreased during the proestrus, reaching a peak 24 hours before estrus in group A and
48 hours before estrus in group B. The VGT then decreased markedly reaching the lowest
value in groups A and B, respectively, 12 and 6 hours after estrus. Although the time ofovulation was only estimated on the basis of a literature review, the matching between thetemporal variations of the VGT values and the predicted time of the peak of estradiol
secretion that ultimately leads to the ovulation processes suggests that the VGT values
represent a potential predictive marker of the ovulatory events.
/C2112014 Elsevier Inc. All rights reserved.
1. Introduction
In recent decades, productivity of the pig industry has
increased dramatically, supported by major investments innew techniques that maximize the genetic potential of the
herds. In reproductive management, arti ficial insemination,
estrus synchronization, and ultrasonography for pregnancy
diagnosis are now widely used. However, determining the
optimal time to inseminate still relies on a relatively sub-
jective evaluation of behavioral and physical signs of the
sow. Because the best fertility results are achieved when
artificial insemination is performed during the 24 hours
before to ovulation [1–3], the development of a technique*Corresponding author. Tel.: ț33 5 61 19 23 06; fax: ț33 5 61 19 39 24.
E-mail address: [anonimizat] (A. Waret-Szkuta).
1Present address: Les Bastides du Sud; 5 Av. Barthélemy Thimonnier;
64140 Lons, France.
Contents lists available at ScienceDirect
Theriogenology
journal homepage: www.theriojournal.com
0093-691X/$ –see front matter /C2112014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.theriogenology.2014.07.017Theriogenology 82 (2014) 1080 –1085

that could accurately predict its occurrence under field
conditions would be of particular value. This would allow
not only a more precise insemination time and, thus, an
increase in the fecundation rate but also a decrease in the
number of inseminations per animal and fewer post-
ovulatory inseminations.
Several studies have tested the applicability of ultraso-
nography to detect or predict the occurrence of ovulation,
but the protocols are usually time-consuming and requiretrained staff [4–6]. Moreover, even if using ultrasonography
has enabled ovulation to be detected, it is not able to pre-
dict it in advance [7,8] . Other authors have found a rela-
tionship between the electrical resistance of the vaginal
mucus and the day of estrus [9,10] , but the results show a
considerable variation among animals and between
different measurement locations in the vagina [10,11] .
More recently, infrared thermography was used to detect
changes in the vulvar skin temperature (VST) of sows during
the periovulatory period [12,13] . A signi ficant decrease in the
vulvar temperature before ovulation was reported, but it
remained unclear whether the temperature peaks observed
occurred exclusively in the vulvar region or as a re flex
response to a rise in body temperature. Moreover, although
several authors have reported temperature fluctuations in
mammalsrelated tothe circadianrhythm [14–16],t h e s ew e r e
not taken into account in these previous studies because the
measurements were only taken at 12-hour intervals.
Infrared thermography is a modern, noninvasive, and
safe technique that measures the temperature of a surface,
based on its emission of infrared radiation. It has been
shown that estrogen administration can induce an increase
in vaginal blood flow measured through a rise in vaginal
thermal conduction [17,18] . The increased local blood flow
linked to rising plasma estrogens is re flected by vulvar
reddening and swelling that have been widely reported as
typical signs of estrus in the sow [19,20] . Infrared ther-
mography has the potential to evaluate these physiological
changes by monitoring the evolution of the VST. The aim of
this study was to detect variations of VST in sows and gilts
during the periovulatory period, to establish a relationship
between these temperature fluctuations and ovulation,
and, ultimately, to evaluate the applicability of this methodto predict ovulation under field conditions.
2. Materials and methods
2.1. Animals and housing
This study took place in the Groupement d ’intérêt
économique (GIE) Villefranche Grand Sud located in Ville-
franche de Rouergue, France, from January 25, 2012, to
February 2, 2012. The experimental group comprised a total
of 36 female sows, of which 30 were multiparous sows with
mean parity of 4.1 /C62.3 and lactation length of
28.3/C60.7 days and 6 were gilts of around 249 days of age.
The females were crossbred Large White /C2Landrace,
housed individually in stalls measuring 2 /C20.60 meters
with controlled temperature and humidity. To synchronize
the gilts, each one received an oral daily dose of 20 mg
(5 mL) of altrenogest (Altresyn, Ceva Santé Animale,
Libourne, France), for 14 days, until 2 days before weaningthe sows. The day when the sows were weaned was
defined as Day 0 of the study.
The room temperature was set at 19/C14C and monitored
by the electronic ventilation system ’s thermostats. How-
ever, the temperature was recorded every time when
thermal imaging was carried out and double-checked with
an independent thermometer placed inside the room.
2.2. Detection of estrus
Starting Day 1, estrus was detected twice daily (morning
and afternoon) by a single person using a 2-year-old Duroc
boar. It was based on the recognition of typical signs such as
standing re flex and ear pricking. The animals detected in
the morning were recorded as starting estrus at 6:00 AM,
whereas the ones detected in the afternoon were recorded
as starting estrus at 12:00 PM.
2.3. Infrared thermography
Temperature measurements of the VST and GST were
done using an infrared thermal camera (Fluke TiR 9 Hz
Thermal Imager, Fluke Corporation, Everett, WA, USA)
every 6 hours at midnight, 6:00 AM,12:00 PM, and 6:00 PM
from Day 1 to Day 7. These thermograms were systemati-
cally obtained at a distance of 1 m from the rear end of the
animal, with a 90/C14incidence. A pink dot was drawn on each
gluteal region, so all the temperature records were ob-
tained from the same location as shown in Figure 1 A. For
each measurement, two sequential thermograms were
recorded, downloaded, and visualized with SmartView 2.0
(Fluke Thermography, Plymouth, MA, USA). The VST re-
cords were estimated by drawing a polygonic area delimi-
ted around the vulva, and the GST was taken in a circle in
each gluteal spot. The recorded VST was de fined by the
mean temperatures from the polygon area, whereas the
GST came from the overall mean of the two gluteal mea-
sures as represented in Figure 1 B. If the image quality of the
first thermogram was poor, it was discarded and the
readings were obtained from the second thermogram.
Because these temperature variables were susceptible
to both exogenous and endogenous factors, a third variable
was created based on the difference between the VST and
the GST. This allowed us to differentiate an increase in the
VST due to body temperature (related to a febrile state for
example), from an increase occurring exclusively in the VST
(which would lead to a subsequent increase in the vulvar –
gluteal temperature (VGT)).
In order to assess the repeatability of this method applied
to the measurement of the VGT, the measurements of the
VST and the GST were carried out five times at 5 minutes
intervals on a sample group of four females at Day 0.
2.4. Statistical analysis
Data were entered in an Excel database (Microsoft Excel,
2010, Microsoft Corporation, USA). Records from two ani-
mals (one sow and one gilt) were discarded as the animals
had not shown any sign of estrus during the study. The
repeatability of the method to measure VGT infrared tem-
perature was estimated by calculating a repeatability indexV.G. Simões et al. / Theriogenology 82 (2014) 1080 –1085 1081

(ri). This r iwas determined following one-way ANOVA [19],
in which the ratio was calculated as follows: variations be-
tween the four animals on which the measurements were
taken and the overall variation (between and within ani-
mals, i.e., at each time of measurement). The VST and GST
were standardized to the time of estrus (t 0¼estrus, t –t0)i n
order to overcome the initial differences from each animal ’s
weaning-to-estrus interval. To investigate the difference
between night (based on the values recorded at midnight
and 6:00 AM) and day (based on the values recorded at
12:00 PM and 6:00 PM) of the VSTand the GST, respectively,
individual animal measures were averaged per respective
group of hours concerned and over the 7 days of the trial and
then a Student ’s t-test was performed. To analyze the vari-
ation of the VST and the GST all along the estrus cycle, re-
cords were averaged per day and for all the animals together,
and one-way ANOVAs were applied. Differences were re-
ported as signi ficant when P value was less than 0.05. At the
same time, to further investigate the signi ficant differences
found at the day level, a detection procedure was applied tothe VGT to enable data at the hour level to be considered for
each animal. The method was developed by Caussinus and
Lyazrhi and the details can be found in [21]. The procedure is
a modi fied Akaike Information Criterion model comparison.
Each model is characterized by a random number of change
points. A change point is a threshold when a jump between
the average temperature of the individuals calculated before
and after is observed. At the end of the procedure, the model
chosen will be the best fit for the data. It will render the
correct number of existing change points (which can be zero
if variability in temperature is seldom due to interindividual
variability). It was first performed on all the data at Day 0 of
the trial as the starting point (n ¼34). It was then repeated
on two subgroups, respectively, for which the animals had
similar weaning-to-estrus interval as determined by the
naked eye. The two subgroups were considered separately
to overcome the differences in weaning-to-estrus intervals
among individuals. The first (group A) consisted of 11 ani-
mals with detected estrus at 6:00 AM of Day 4. The second
(group B) consisted of seven animals with detected estrus at
6:00 AM of Day 5. Statistical analyses were conducted usingSPSS 20.0 for Windows (IBM Corp., USA) and R software,
version 2.13 (R Foundation for Statistical Computing, 2013).
3. Results
The room temperature varied not only between day
(average value at 12:00 PM and 6:00 PM) and night
(average value at midnight and 6:00 AM), with the highest
temperatures recorded at 6:00 PM (mean /C6SD: 17.8 /C61.6
/C14C) and the lowest at 6:00 AM (16.1 /C62.2/C14C), but also
during the course of the trial, decreasing from Day 1
(19.3/C61.2/C14C) to Day 7 (15.0 /C60.8/C14C).
An r iof 0.926 was determined for the infrared VGT
measurement method. A signi ficant daily variation pattern
in the VST was found between day and night (P <0.001),
with the highest temperatures obtained at 6:00 PM
(mean /C6SD: 32.5 /C62/C14C) and the lowest at 6:00 AM
(29.9/C62.6/C14C). The same daily variation pattern was found
for the GST (P <0.001), ranging from 30.7 /C61.9/C14C at 6:00
PM to 27 /C62.6/C14C at 6:00 AM.
The VST remained relatively stable during the proestrus
(mean /C6SD: 32.3 /C61.8/C14C), but not signi ficantly (P ¼0.86),
but starting to decrease with the estrus (P <0.001). On Day
1 of estrus, the mean VST /C6SD was 31.7 /C62.4/C14C. On Day 2, it
was 31.0 /C62.4/C14C, and on Day 3, it was 29.6 /C63.1/C14C. The
GST did not show any signi ficant variation throughout the
trial (P ¼0.47).
The analysis of the VGT values obtained from all the
animals showed five change-points at /C036 hours,
/C024 hours, ț6 hours, ț18 hours, and ț24 hours to estrus
(Fig. 2 ). The VGT remained initially stable, but then started
to decrease 36 hours to 24 hours before estrus, reaching the
lowest value 12 to 18 hours after estrus and then stabiliz-
ing, that is to say increasing very slightly.
In group A, one change point was found 6 hours after
estrus. The VGT increased during the proestrus, reaching a
maximum mean value of 3.8 /C61.9/C14C 24 hours before
estrus, after which it started decreasing markedly to reach a
minimum value (0.5 /C61.5/C14C) 12 hours after estrus ( Fig. 3 ).
In group B, one change point was found 24 hours before the
estrus. The VGT increased during the proestrus, reaching a
Fig. 1. Comparison between a normal and a thermal image of the posterior end of a sow. (A) White light image with evidence of the auxiliary spots from where
the gluteal temperature was obtained. (B) Gray scale thermogram showing the method used to obtain the temperature records. (For interpretation of th e ref-
erences to color in this figure, the reader is referred to the web version of this article.)V.G. Simões et al. / Theriogenology 82 (2014) 1080 –1085 1082

maximum mean value of 5.3 /C62.4/C14C 48 hours before
estrus, after which it started decreasing to reach the lowest
value 6 hours after estrus (1.1 /C60.9/C14C,Fig. 4 ).
4. Discussion
This study enabled us to assess the VST of sows and to
follow the dynamic during the periovulatory period. The r i
obtained corresponds to a highly repeatable method [22]
and con firms the applicability of infrared thermography
for evaluation of skin temperature in sows.
The variation in the VSTand GST throughout the day with
the lower temperatures obtained in the early morning and
the higher temperatures in the early evening are inagreement with the results from another study that moni-
tored the vaginal temperature using internal temperature
transmitters [23]. These daily fluctuations have been widely
described in the literature as a result of the circadian
rhythm, an endogenous thermal regulation mechanism
responsible for a temperature fall during the night period
[14–16]. The fluctuations in the room temperature also
contributed to emphasize the magnitude of these variations
because the highest and lowest skin temperatures recorded,
respectively, at 6:00 PM and 6:00 AM matched with the
maximum and minimum peaks of the room temperatures
measured at the same periods of the day.
Surprisingly, the VST did not change signi ficantly during
the proestrus. In a similar study, a signi ficant increase was
reported in the VST as the estrus approached [12] due to the
vulvar swelling and reddening. Failure in detecting this
increase here may be explained by the steady decrease in
the room temperature throughout the study, which
masked the expected increase in the VST. Assuming that
the decrease in the room temperature affected equally the
VST and the GST, the analysis of the VGT had a special
relevance, because it re flected the real trend of the VST
without the in fluence of the room temperature.
The VGT analysis of the entire experimental group
revealed a high number of change points, showing, as ex-
pected, the different temperature pro files among animals
related to the asynchrony of their estrous cycles or at least
to the different duration from weaning to estrus. Indeed,
assuming that these temperature variations are related to
the estrous cycle [12,13] and because the duration of estrus
varied widely between animals, ranging from 36 to
72 hours, a considerable heterogeneity in temperature
profiles was expected when evaluating animals with
different estrus durations. Groups A and B included animals
Fig. 2. Variation in vulvar –gluteal temperature (VGT) and 4-points average
from all animals. The gray shaded area represents the SD of the VGT. TheVGT started decreasing 24 to 36 hours before estrus (OE) and stabilized 12 to
28 hours after OE. Five change points were detected.
Fig. 3. Variation in vulvar –gluteal temperature (VGT) in group A (n ¼11).
The gray shaded area around the VGT represents its standard deviation (SD)and the gray gradient the standing estrus. A change point is detected 6 hoursafter estrus, with a maximum (M) temperature peak obtained 24 beforeestrus and a minimum (m) temperature peak 12 hours after estrus.
Fig. 4. Variation in vulvar –gluteal temperature (VGT) in group B (n ¼7). The
gray shaded area around the VGT represents its SD and the gray gradient thestanding estrus. A change point is detected 24 hours before estrus, with amaximum (M) temperature peak obtained 48 hours before estrus and aminimum (m) temperature peak 6 hours after estrus.V.G. Simões et al. / Theriogenology 82 (2014) 1080 –1085 1083

with the same weaning-to-estrus interval and conse-
quently with a more homogenous estrus duration. In both
the groups, only one change point was found, respectively,
6 and 24 hours before estrus. Although the interpretation of
the timing of these change points is limited according to
biological principles, it proves from a statistical point of
view the existence of a noticeable deviation in the vulvar
temperature relative to the body temperature.
The highest differential between the vulvar and the body
temperature, represented by the VGT, was found 24 and
48 hours before estrus in groups A and B, respectively. These
peaks coincided with the plateau in the swelling and
reddening of the vulva usually occurring 24 to 36 hours
before estrus [24]. Although it is recognized that these
physiological changes are due to an increased blood flow in
the vulvar tissues, the underlying mechanisms promoting
these changes are still not absolutely clear. A recent study
[25] used infrared thermography to monitor the estrous
cycle in mares and reported an increase in the vulvar tem-
perature during the follicular growth and a decrease of the
same temperature during the establishment of corpora
lutea. The authors concluded that these temperature vari-
ations were probably a result of the high levels of estradiol
produced by the preovulatory follicles. In primates, admin-
istration of estradiol in ovarectomized females resulted in
reddening and swelling of the genital region, mimicking the
signs observed during estrus [26]. In pigs, the estradiol peak
occurs 24 to 48 hours before estrus [27]. Our temperature
peaks were found within the same intervals, which suggests
that the changes we observed in the vulvar tissues were
most probably also induced by this hormone.
Ovulation in pigs results from a series of well-de fined
events: the developing preovulatory follicles secrete
increasing amounts of estradiol, culminating in a peak 24 to
48 hours before estrus. This estradiol peak triggers the
single preovulatory surge of LH that initiates the rupture of
the follicle wall, culminating in ovulation [28]. According to
the results of different studies, the occurrence of these
events is fairly constant in time, ranging from 10.6 /C61.6 to
12.6/C62.4 hours between the estradiol and LH peaks and
from 30.6 /C61.3 to 37.1 /C62.4 hours from the LH peak to
ovulation [29,30] . Because the variations observed in the
VGT are apparently related to the levels of estradiol and
given the role of this hormone in triggering the ovulation
mechanisms, it can be postulated that vulvar temperature
has the potential to be used as a predictive marker of the
ovulatory events.
Infrared thermography proved to be, however, highly
sensitive to changes in the environmental conditions, as
already discussed by other authors [31,32] . Factors such as
airflow, moisture, fluctuations in the environmental tem-
perature, level of physical activity, and animal ’s stance
before the measurement can induce a considerable varia-
tion in these readings, which may limit the applicability of
this technology under field conditions where these factors
are dif ficult to control.
4.1. Conclusions
Based on our findings and on the literature data pre-
sented, it appears that (1) vulvar temperature changessignificantly during the periovulatory period, (2) the tem-
perature peak found is due to the high levels of estradiol,
and (3) vulvar temperature has the potential to be used as a
potential predictive marker of ovulation. Because the
occurrence of ovulation could not be precisely determined,
the accuracy of this technique remains to be studied
further. Hormone assays (FSH, LH, progesterone, and es-
trogen) and ovary ultrasonography should be performed
together with infrared thermography in order to evaluatethe relationship between vulvar temperature variations
and hormone levels, and also to determine with greater
accuracy the time of ovulation. The use of this technology
under field conditions should be carefully considered due
to its high sensitivity to environmental fluctuations.
Acknowledgments
We thank all the staff from the Bernussou Training
Center for having hosted this study in their facilities.
References
[1]Soede NM, Wetzels CCH, Zondag W, de Koning MAI, Kemp B. Effects
of time of insemination relative to ovulation, as determined by ul-
trasonography, on fertilization rate and accessory sperm count in
sows. J Reprod Fertil 1995;104:99 –106.
[2]Nissen AK, Soede NM, Hyttel P, Schmidt M, D ’Hoore L. The in fluence
of time of insemination relative to time of ovulation on farrowingfrequency and litter size in sows as investigated by ultrasonography.
Theriogenology 1997;47:1571 –82.
[3] Cassar G. Proper timing for sow ovulation. Proceedings 25th Cen-
tralia Swine Research Update; 2006 Jan 25. Kirkton, Canada.
[4]Soede M, Helmond FA, Kemp B. Periovulatory pro files of oestradiol,
LH and progesterone in relation to oestrus and embryo mortality in
multiparous sows using transrectal ultrasonography to detect
ovulation. J Reprod Fertil 1994;101:633 –41.
[5]Knox RV, Althouse GC. Visualizing the reproductive tract of the fe-
male pig using real-time ultrasonography. J Swine Health Prod1999;7:207 –15.
[6]Kauffold J, Althouse GC. An update on the use of B-mode ultraso-
nography in female pig reproduction. Theriogenology 2007;67:901 –
11.
[7]Soede NM, Hazeleger W, Kemp B. Follicle size and the process of
ovulation in sows as studied with ultrasound. Reprod Domest Anim
1998;33:239 –44.
[8]Waberski D, Kunz-Schmidt A, Neto GB, Richter L, Weitze KF. Real-
time ultrasound diagnosis of ovulation and ovarian cysts in sows
and its impact on arti ficial insemination ef ficiency. J Anim Sci 2000;
77:1 –8.
[9]Rezac P, Kukla R, Poschl M. Effect of sow parity on vaginal electrical
impedance. Anim Reprod Sci 2002;72:223 –34.
[10] Stokhof S, Soede NM, Kemp B. Vaginal mucus conductivity as
measured by the Walsmeta MKIV does not accurately predict the
moment of ovulation or the optimum time for insemination in
sows. Anim Reprod Sci 1996;41:305 –10.
[11] Rezac P, Poschl M, Krivanek I. Effect of probe location on changes in
vaginal electrical impedance during the porcine oestrus cycle.Theriogenology 2003;59:1325 –34.
[12] Scolari SC, Clark SG, Knox RV, Tamassia M. Vulvar skin temperature
changes signi ficantly during estrus in swine as determined by dig-
ital infrared thermography. J Swine Health Prod 2011;19:151 –5.
[13] Luño V, Gil L, Jerez RA, Malo C, González N, Grandía J, et al. Deter-
mination of ovulation time in sows based on skin temperature and
genital electrical resistance changes. Vet Rec 2013;172:579 .
[14] Kräuchi K. How is the circadian rhythm of core body temperature
regulated? Clin Auton Res 2002;12:147 –9
.
[15] Kräuchi K. The human sleep –wake cycle reconsidered from a
thermoregulatory point of view. Physiol Behav 2007;90:236 –45.
[16] Weinert D. Circadian temperature variation and ageing. Ageing Res
Rev 2010;9:51 –60.
[17] Abrams RM, Stolwijk JAJ. Heat flow device for vaginal blood flow
studies. J Appl Physiol 1972;33:144 –6.V.G. Simões et al. / Theriogenology 82 (2014) 1080 –1085 1084

[18] Abrams RM, Thatcher WW, Bazer FW, Wilcox CJ. Effect of estradiol-
17beta on vaginal thermal conductance in cattle. J Dairy Sci 1973;
56:1058 –62.
[19] Langendijk P, van den Brand H, Soede NM, Kemp B. Effect of boar
contact on follicular development and on estrus expression after
weaning in primiparous sows. Theriogenology 2000;54:1295 –303.
[20] Sterning M. Oestrous symptoms in primiparous sows. 2. Factors
influencing the duration and intensity of external oestrous symp-
toms. Anim Reprod Sci 1995;40:165 –74.
[21] Caussinus H, Lyazrhi F. Choosing a linear model with a random
number of change-points and outliers. Ann Inst Statist Math 1997;
49:761 –75.
[22] Measey GJ, Silva JM, Di Bernardo M. Testing for repeatability in
measurements of length and mass in Chthonerpeton indistinctum
(Amphibia: Gymnophiona); including a novel method of calculating
total length of live caecilians. Herpetol Rev 2002;34:35 –9.
[23] Soede NM, Hazeleger W, Broos J, Kemp B. Vaginal temperature is
not related to the time of ovulation in sows. Anim Reprod Sci 1997;
47:245 –52.
[24] Worwood D. Swine arti ficial insemination for beginners: heat
detection [Internet]. Emery County: Utah State University; 2007[cited 2012 Apr 23, 2012]. Available from http://extension.usu.edu/
files/publications/publication/AG_Swine_2007-02.pdf .[25] Stelletta C, Gianesella M, Vencato J, Fiore E, Morgante M. Thermo-
graphic applications in veterinary medicine. In: Prakash RV, editor.
Infrared thermography. Rijeka: InTech; 2012. p. 117 –40.
[26] Dixson AF. Observation on the evolution and behavioral signi ficance
of“sexual skin ”in female primates. In: Rosenblatt JS, editor. Adv Stud
Behav, vol. 13. New York: Academic Press, Inc; 1983. p. 63 –101.
[27] Guthrie HD, Henricks DM, Handlin DL. Plasma estrogen, proges-
terone and luteinizing hormone prior to estrus and during early
pregnancy in pigs. Endocrinology 1972;91:675 –9.
[28] Whittemore CT, Kyriazakis I. Whittemore ’s science and practice of
pig production. Third edition. Oxford: Blackwell; 2006 .
[29] Lang A, Brandt Y, Madej A, Einarsson S. In fluence of simulated stress
during standing oestrus on ovulation and hormonal pro file in the
sow. Reprod Domest Anim 2004;39:255 .
[30] Mburu JN, Einarsson S, Dalin AM, Rodriguez-Martinez H. Ovulation
as determined by transrectal ultrasonography in multiparous sows:relationship with oestrous symptoms and hormonal pro files. J Vet
Med A 1995;42:285
–92.
[31] Sykes DJ, Couvillion JS, Cromiak A, Bowers S, Schenck E,
Crenshaw M, et al. The use of digital infrared thermal imaging to
detect estrus in gilts. Theriogenology 2012;78:147 –52.
[32] Cravello B, Ferri A. Relationships between skin properties and
environmental parameters. Skin Res Technol 2008;14:180 –6.V.G. Simões et al. / Theriogenology 82 (2014) 1080 –1085 1085