PROJECT RESULTS

Optimization of Lighting Conditions in Dairy Barns

Applicant: 

Dr. Alma Kennedy

Department of Animal Science

University of Manitoba

Winnipeg, Manitoba  R3T 2N2  Canada

Table of Contents:

• Experiment 1
• Experiment 2
• Conclusions and Implications

ARDI Project:

#00-389

Total Approved:

Date Approved:

Project Status:

Completed August, 2004

In the past five years it has become apparent that dairy heifer and cow function can be modified by alterations in the indoor lighting conditions and recommendations have been made regarding light intensity and duration of light exposure during the day.  Some of these recommendations, which are meant to increase milk production, can result in inconvenience for the dairy producer and an increase in the cost of barn lighting.  As well, unanswered questions regarding barn lighting conditions still remain.  The studies described herein were designed to provide answers which would help to ensure that, with respect to dairy barn lighting conditions, convenience for the dairy producer is maximized, lighting cost is minimized and milk production is not compromised in any way.

In Experiment 1 we examined the effect of various intensities of dim light on dairy heifer physiology and in Experiment 2 we examined the effect of various day lengths on milk production of dairy cows.  The results have been previously published in the M.Sc. Thesis of Pushpa Mutharamalingam.

Experiment 1 - Plasma Melatonin and IGF-1 Responses to Dim Light at Night in Prepubertal Dairy Heifers

Background:

Melatonin hormone is produced by the pineal gland of the brain only at night.  This results in the very regular occurrence of high levels of melatonin in the blood (a high plateau in cattle) at night and levels of essentially zero during the day for a wide variety of species.  This daily rhythm in blood melatonin is thought to direct other daily rhythms in the body and is thought to be important for good animal health.  Light intensities of 150 to 200 Lux are recommended during daytime as studies (reviewed by Dahl et al. 2000) have shown that dairy cattle respond to such intensities with an increase in milk production when the day length is extended to 16 to 18 hours.

Through previous Manitoba Hydro and ARDI funding, we (Lawson and Kennedy 2001) have previously shown that a light intensity of 50 Lux causes a reduction in night melatonin level in dairy heifers but lower intensities of light were not examined.  Dahl et al. (2000) have emphasized that 6 to 8 hours of complete darkness is important for dairy cows to ensure the continuation of natural circadian rhythms.  Manitoba Hydro also makes this recommendation when advising dairy producers on lighting retrofits or new barn lighting.  At this point we can only say that the use of a 50 Lux light intensity at night will likely shorten the length of the night perceived by dairy cows and so may be bad with respect to milk production and cow health.  Dahl et al. (2000) recommends only red light be used at night, if a night light is needed.  Most species, including cattle, do not respond to light in the red end of the visible spectrum.  This recommendation has generally not been accepted in the industry as it is not practical.

Many dairy farmers still provide either dim light (less than 100 Lux) or bright light (greater than 100 Lux) at night to promote feed consumption at night and to facilitate health checks during the night.  Although research has shown that bright night light does not promote feeding at night (Tanida et al. 1984) and that it may actually decrease feed intake (Peters et al. 1981), the practice continues.  Because there have never been studies showing negative production effects of dim light at night, the practice of the use of dim light is considered acceptable by many producers.

Animal species differ dramatically in their sensitivity to light with the highest sensitivity being found, logically, in nocturnal species.  Theoretically, for every species, there is an intensity of light or limit of sensitivity, below which the animal will demonstrate no physiological response to the light.  This level in sheep (Arendt and Ravault 1988) is very low (1.0 Lux or lower) but sensitivity to dim light in cattle has not been previously studied using very dim levels of light.  Thus, we do not know what the "safe" intensity of dim nightlight for cattle is.  Our previous research (Lawson and Kennedy 2001) suggested that the intensity was below 50 Lux.

Objective:

To determine if low light intensities have a physiological effect in dairy heifers.

Experimental Approach:

Plasma melatonin levels were measured in dairy heifers exposed at night (8 hours) for a short (1 day) and long (14 days) time to no dim light (0 Lux), two very low intensities of dim light (5 and 10 Lux) or an intensity of dim light (50 Lux) previously shown to cause a physiological effect.  The 5 and 10 Lux intensities were chosen as it would be possible for dairy producers to check cows at these intensities of light. An intensity of light less than 5 Lux would be too dim for practical use.

Methods:

Twelve dairy heifers were exposed to 0, 5, 10 or 50 Lux (0, 0.5, 1.0 or 5 fc) light intensity during the 8 hour nighttime period for 14 consecutive days using a Latin Square Design which made it possible for every heifer to receive every treatment once.  Eric Witkowski, lighting engineer, Manitoba Hydro, designed the incandescent and fluorescent lighting schemes for four rooms in the Animal Science Research Unit at the University of Manitoba for this experiment.  Fluorescent fixtures were used for lighting during the day and incandescent fixtures were used for experimental lighting during the night.  Bulbs of a variety of wattages were used to achieve uniformity in the dim light intensities required throughout the animal holding areas.  Blood samples were collected via jugular catheter during the last hour of the light period and at 1, 2, 4 and 8 hours of the night on days 1, 4 and 14 of the experiment.  The heifers were exposed to a light intensity of 200 Lux (20 fc) during the 16 hour days.

Results and Discussion:

Statistical analysis revealed that there was a significant Treatment X Hour effect which is shown in Figure 1 where means represent a particular hour averaged over days 1, 4 and 14 of the experiment.  Of the three light treatments, only the 50 Lux treatment affected plasma melatonin level and this effect had disappeared by the 4th hour of the night.  Plasma melatonin only rose to 50% of the normal night level when the 50 Lux light was present at night.  Although the negative effect of 50 Lux on melatonin disappeared by 4 hours into the night, because melatonin analysis was done only on samples collected at 2 and 4 hours of the night, the negative effect may have lasted for 3 hours or more.  This inhibitory effect of 50 Lux was equally dramatic at l, 4 and 14 days of treatment.  Because the ability of 50 Lux to depress melatonin at night occurred consistently throughout the 14 day experiment, we feel confident that the heifers did not become insensitive to 50 Lux over time.  This is important in that one cannot assume that animals will "get over" the negative effect of 50 Lux on plasma melatonin levels.

It is necessary to question whether there would be functional consequences to a 50% reduction in plasma melatonin at night which was caused by the 50 Lux treatment.  Although this question cannot be answered at this time, related information in sheep suggests that the height of the nighttime melatonin plateau is of practical importance.  As found in sheep (Hotter and Chemineau 2001) we have observed a high degree of variability among dairy heifers for the normal night level of plasma melatonin.  Some heifers consistently have a high plateau while other consistently have a plateau ½ to ¼ the highest levels found.  In sheep, nighttime melatonin plateau levels are negatively correlated to fertility (Hotter and Chemineau 2001).  It would be very interesting to examine this in cattle but as of yet this has not been attempted.

Melatonin values at night for the 5 and 10 Lux treatments were not different from values found during complete darkness (0 Lux) (Figure 1).  Thus, the 0, 5 and 10 Lux treated heifers all showed the normal rise in plasma melatonin at night.  Unfortunately, we were limited to having 4 treatments in this experiment and so were unable to test the effects of the intermediate light intensities of 20, 30 and 40 Lux.  Results found on day 4 and 14 were identical to results found on day 1 indicating that the heifer sensitivity to 5 and 10 Lux does not develop over time.

Our results indicate that dairy heifers are not sensitive to 5 or 10 Lux light during the night but respond to 50 Lux for greater than 2 but less than 4 hours per night.  If light with an intensity of 50 Lux is used at night in dairy barns, cows will receive less "physiological" hours of darkness than intended, and possibly needed, for good health.  The heifers showed no tendency to develop either increased or decreased sensitivity to dim light over the course of the 14 day experimental period.  Dahl et al. (2000) has indicated that 6-8 hours of darkness is important for dairy cows in order to ensure the continuation of natural circadian rhythms.  Currently there are dairy producers who provide light at 50 Lux or higher at night.  In such cases, the presence of light at night may be reducing the length of night to shorter than that recommended.  A recommendation for the use of 5 or 10 Lux intensity at night in dairy barns would ensure that light at night does not compromise cow productivity and health.

Conclusions:

Current recommendations are that it is important for dairy cattle to be exposed to a certain number of hours of complete darkness each night (at least 6 hours according to industry experts) but many dairy producers leave lights on in the barn at night for a variety of practical reasons.  Our results indicate that a night light of 50 Lux has a physiological impact on cattle but intensities of 5 and 10 Lux do not.  We were limited to having 4 treatments and so were unable to test the effects of the intermediate light intensities of 20, 30 and 40 Lux.  This study is the first in the world to study, under controlled conditions, the effects of very low light intensities on cattle physiology.  The results suggest that farmers should be very cautious when using light intensities greater than 10 Lux during the night.  Light of 10 Lux intensity is bright enough for humans, and likely cattle, to see fairly well if the eyes are allowed a minute or two to adapt.  Thus, use of dim light up to 10 Lux in intensity is a practical and safe alternative for dairy producers who wish to have light in their free-stall or tie-stall barns during the night.

Experiment 2 - Milk Production, Milk Composition and Plasma IGF-1 in Dairy Cows Exposed to 14, 16, 18, and 20 h Photoperiods

Background:

Dahl et al. (2000) reviewed a large number of studies where exposure of dairy cows to 16 to 18 hour day length increases milk production significantly compared to 13 or lower hours of day length.  They recommend that day length be no longer than 16 to 18 hours (length of total darkness at night no shorter than 6 to 8 hours) in order to ensure the continuation of natural circadian rhythms.  Manitoba Hydro also makes this recommendation when advising dairy producers on lighting retrofits or new barn lighting.  However, only indirect evidence was used to conclude that day lengths longer than 18 can be deleterious to milk production and that 14 hours is too short of a day to generate this effect.  Having a day length as short as 14 hours or longer than 18 hours could be very convenient and/or economical for dairy producers as they would have more flexibility in management in terms of when lights must be turned off or on.  This could lead to a saving in hydro bills and bulb costs (use of a 14 hour day) or result in improvements in ability to check the cows at night (if it is safe to have lights on for 20 hours a day).  Based on solid evidence, exposure of dairy cows to light for 24 hours a day is not recommended.

Twenty-four hour a day light exposure resulted in reduced feed intake of dairy cows (Peters et al. 1981).  Because milk production using 14 or 20 hour day lengths has never been directly compared to that using 16 or 18 h day lengths, a study was done to make this comparison using lactating dairy cows.

Objectives:

To compare milk yield, milk composition, body weight, body condition score, feed intake and plasma IGF-1 levels in lactating dairy cows exposed to day lengths of 14 16, 18, and 20 hours.

Methods:

Milk composition, body weight, body condition score, feed intake and plasma IGF-1 levels were measured in dairy cows exposed to day lengths of 14, 16, 18 and 20 hours at the Glenlea Dairy Facility.  Twelve cows (4 primiparous and 8 multiparous, 109 ± 14 d in milk at the start of the trial) were used but two were eliminated from the study due to poor health.  Breeding of the cows was synchronized so that each would be in early lactation at the beginning of the study which was 168 days in length.  The cows weighed 695.6 ± 57.8 kg (mean ± SD).  Cows had unlimited access to feed and water and were fed 1 kg hd-1 d-1 of alfalfa hay (Table 1). The TMR was balanced for the production of 35 kg of milk/day/cow.  Cows were milked twice daily (04:30 h and 16:30 h).

Table 1.  Composition of the Diet for Experiment 2 

z Containing (g kg-1): Alfalfa silage (315); Corn silage (320); Dairy 26 (260); DPE (95); Sunflower Seed (10).

y Energy content expressed as NEL (Mcal kg-1) (Cornell Nutrition Conference 1979 and Penn State Lab Nutrition 1978).

Each cow was given each treatment for 6 weeks using a Latin Square Design.  Administration of 4 lengths of day in the Glenlea tie-stall barn was achieved by the use of polyethylene curtains that separated cows, in treatment groups of three, from all other cows.  Separate light controls and fluorescent fixtures were installed for each of the 4 curtained areas to achieve a light intensity of approximately 250 Lux in each area during the day. The curtains were all raised in the morning to allow milking and lowered in the evening so that some cows could receive different day lengths.

Milk production was measure daily.  Feed intake was measured over the final 2 days of weeks 5 and 6 of each treatment.  Milk samples were collected during weeks 5 and 6 for analysis of milk fat, protein, SNF and somatic cell count.  The daily amount of TMR given to each group of cows was recorded during periods 2, 3 and 4.  Weigh backs for each group were recorded weekly during weeks 1 to 4 and daily during week 5 and week 6.  Representative samples of the weigh backs and the TMR were collected.  Core samples of the hay were taken from each new bale fed during experimental periods 2, 3 and 4.  Daily samples of the TMR and weigh backs were pooled for weeks 5 and 6 separately and hay samples from all periods were pooled for proximate analysis.

Blood samples were collected by tail vein puncture on day 7 of week 3 and week 6 of each period using 10 ml heparinized vacutainer tubes and 20 G needles for measurement of IGF-1 level.  Body weight, body condition score were conducted during the 6th week of each treatment.

Two cows were removed from trial, one due to an udder injury and subsequent mastitis, and another due to a dramatic drop in milk yield during period 2 of the trail.  Data for these cows were not included in the statistical analysis.

Milk samples were preserved with 2-bromo-2-nitropropane-1,3 diol and stored at 4°C until analyzed for composition using near infrared spectroscopy (NIR) with the Milk-O-Scan 303AB (Floss Electric, Hillerod, Denmark) at the laboratory of the Manitoba Milk Producers (Winnipeg, MB).

Dry matter content of both the feed and weighbacks was determined by drying at 60°C for 48 h (AOAC, 1990).  All feed samples were analyzed for crude protein (Mixed Catalyst Kjeldahl procedure, AOAC, 1990), NDF (National Forage Testing Association, 1993) and ADF (AOAC, 1990).  Ca, P, K, Mg and Na were measured using inductively coupled plasma emission spectroscopy (AOAC, 1990).  Feed efficiency (kg of milk/kg of dry matter intake) was calculated for individual cows.

Plasma was collected after blood centrifugation (3000 x g for 30 minutes) and stored at - 20°C until assayed for IGF-1 (Kerr et al. 1990) at the Western College of Veterinary Medicine, Saskatoon, SK.

Analysis of variance was carried out to determine the significance of treatment on the milk yield, milk composition, body weight, body condition scoring, feed intake and plasma IGF-1 level.  Statistical analysis of the data was performed using Mixed Model procedure of SAS (SAS Institute, Cary, NY, 1999).  Initially, data for weeks 5 and 6 of each period were analyzed separately, but was then averaged prior to analysis, as the effect of week was not significant (P>0.05).  Cows were at 192 ± 15 (mean ± SD) days in milk by the beginning of period 4.  It was considered that light treatments might not be as effective on milk yield in late lactation as in mid- lactation.  For this reason, statistical analysis was performed with and without period 4 in the data set.  In this analysis, because dry matter intake was available for only periods 2 and 3, intake was not assessed statistically but arithmetic means were calculated.  All effects were tested against the residual error.  Initial milk yield (week 1) was used as a covariate in the analysis, but was non-significant.  Parity, days of pregnancy and days in milk were used as covariates in the model, but were dropped from subsequent analysis because treatment effects in the covariate analysis remained non-significant even though the covariates were found to be significant (P = 0.05).

Results and Discussion:

Body weight and body condition scoring did not differ among treatments throughout the experiment.  There was no effect of day length on milk yield, dry matter intake and feed efficiency whether results of 4 periods (Table 2) or 3 periods (Table 3) of the experiment were analyzed.  Thus, for these parameters, inclusion of late lactation data (period 4, Table 3) made no difference.  It must be noted that the variation among treatments for milk yield was very high (Tables 2 and 3).  Because of this variation, at least a 25% difference among treatments in milk yield would be required to be significant.  In a review, Dahl et al. (2000) indicated that the milk response to long photoperiod (16 h) is sometimes only a 6% increase.  Such an increase would be very difficult to demonstrate using the cows of the Glenlea dairy herd.  At this point we can only conclude that none of the day length treatments had a large effect on milk production.  The variability in our results suggests a number of possible problems.  To find moderate or low milk production differences, we would need to use a large number of cows - more than the 12 used in this study.  It is also possible that the lack of treatment differences related to the 6­-week treatment period which may have been too short to allow for a full photoperiodic response to develop.  However, others (Dahl et al. 1997) have seen a response in 4 weeks.  Also, all other studies (reviewed by Dahl et al. 2000) used cows which had been exposed to a short natural day length (<12 hours) or a short artificial photoperiod (13 hours) prior to their studies.  Cows were exposed to an artificial 18 hour day prior to our study.

Table 2.  Effect of Photoperiod on Milk Yield, Milk Composition and Plasma IGF-1 Levels in Dairy Cows (Least Square Means) When All Four Periods of the Experiment Were Included

z SEM = pooled standard error of the mean

y kg of milk /kg of feed

Table 3.  Effect of Photoperiod on Milk Yield, Milk Composition and Plasma IGF-1 Levels in Dairy Cows (Least Square Means) When Only the First Three Periods of the Experiment Were Included

z SEM = pooled standard error of the mean

y kg of milk /kg of feed

Plasma IGF-1 concentration did not differ among treatments (Tables 2 and 3).  The level of variation was similar to that found in other studies where changes in day length did affect the plasma level of IGF-1.  It is thought that long photoperiod increases milk production through the action of IGF-1 hormone.  Spicer et al. (1994) found an increase in the concentration of blood IGF-1 hormone in heifers exposed to 16 hour (159 ng ml-1) and 18 hour (133 ng ml-1 ) compared to 8 hour (111 ng ml-1) day lengths.  A 24 hour day length resulted in the same IGF-1 level as the 8 hour day length.  Dahl et al. (1997) found a rise in dairy cow plasma IGF-1 after only 4 weeks of exposure to an 18 hour photoperiod compared to natural photoperiod of less than 13 hours per day.  Possibly an effect was not found in our experiment because the shortest day length that we studied was 14 hours rather than the 8 hour day used by Spicer et al. (1994) or the natural short day photoperiod used by Dahl et al. (1997).  Day lengths shorter than 14 hours were not examined in the present study as they would not be representative of a practical indoor dairy barn environment.  Our inability to demonstrate a difference in IGF-1 among our treatment groups is the strongest evidence we have that day lengths of 14 to 20 hours are all perceived as long days by dairy cows.

When Period 4 was excluded from the data analysis (Table 3) we found that total milk fat production tended (P = 0.09) to be elevated for cows exposed to 18 hours of light per day (1.2 Kg d-') compared to cows exposed to 14 (1.0 Kg d-1) or 16(1.0 Kg d-1) hours of light per day and results for 20 hours of light per day (1.1 Kg d-1) were intermediate.  The tendency for a rise in milk fat with longer day length is surprising as others have found no effect of long days (Phillips and Schofield 1989 - experiment 1 and Dahl et al. 1997) or a decrease in milk fat (Stanisiewski et al. 1985 and Phillips and Schofield 1989 - experiment 2) with long days.  Our results may relate to the effect of long days on cow behaviour found by Phillips and Schofield (1989) where cows exposed to long photoperiod spent 20% more time lying down.  Increased time spent lying would reduce maintenance cost and thus would make more feed energy available for body fat or milk fat production.

Conclusions:

There was no evidence in our study for a difference in milk yield, feed intake or feed efficiency due to day lengths of 14, 16, 18 and 20 hours.  However, high levels of variation made it impossible to detect small or moderate differences.

Total milk fat produced per day tended to be higher for cows exposed to 18 hours of light per day.

No difference in plasma IGF-1 level was found in cows exposed to day lengths of 14, 16, 18 and 20 hours.

Overall Conclusions and Implications:

Our results indicate that dairy heifers are not sensitive to 5 or 10 Lux light intensities during the night but respond to 50 Lux for greater than 2 but less than 4 hours per night.  If light with an intensity of 50 Lux is used at night in dairy barns, cows will receive less "physiological" hours of darkness than intended, and possibly needed, for good health.  The heifers showed no tendency to develop either increased or decreased sensitivity to dim light over the course of the 13 day experimental period.  Currently there are dairy producers who provide light at 50 Lux or higher at night.  In such cases, the presence of light at night may be reducing the length of night to shorter than that recommended.  A recommendation for the use of 5 or 10 Lux intensity at night in dairy barns would ensure that light at night does not compromise cow productivity and health.

Milk production and IGF-1 levels were similar for cows exposed to 14 h to 20 h of light per day.  Because the variation among treatments for milk yield was very high, at least a 25% difference among treatments in milk yield would be required to be significant.  At this point we can only conclude that none of the day length treatments had a large effect on milk production.  A treatment difference may also have been missed if the treatment duration of 6-weeks was too short in duration to allow for a full photoperiodic response to develop.  Variation among treatments for plasma IGF-1 level was not excessively large yet no differences were found among treatments for this hormone which is thought to mediate the effect of light on milk production.  This is our strongest evidence that day lengths of 14 to 20 hours are all perceived as long days by dairy cows.  When Period 4 was excluded from the data analysis, we found that total milk fat production tended to be elevated for cows exposed to 18 hours of light per day compared to cows exposed to 14 or 16 hours of light per day and results for 20 hours of light per day were intermediate.  Our results may relate to the effect of long days on cow behaviour as it has been shown by others that cows exposed to long photoperiod spend 20% more time lying down.  Increased time spent lying with a day length of 18 hours would reduce maintenance cost and thus would make more feed energy available for body fat or milk fat production.

Acknowledgements:

The authors are grateful to the Agri-Food Research and Development Initiative (ARDI) and Manitoba Hydro which made it possible to conduct these studies.  Professors K. Plaizier and G. Crow assisted with experimental planning and statistical analysis.  Staff at the Animal Science Research Unit and Glenlea Dairy Barn provided expert animal care and assisted with data collection.  Eric Witkowski and Ray Boris of Manitoba Hydro provided technical guidance for both experiments.

Each of the following people made a great effort to ensure that the experiments were conducted properly and in the process also received training in the field of animal experimentation.  We wish them best wishes in their future endeavors.

• The M.Sc. student, Pushpa Muthuramalingam, received her degree at the University of Manitoba spring convocation in 2004.  Currently she is employed as a research associate at a private Winnipeg molecular biology lab.
• A summer student, Kristen Ficzycz, is currently attending the Veterinary College in Saskatoon.
• A research assistant, Deanne Fulawka, is currently employed in the Department of Animal Science.
• Dr. Rob Berry, a research associate who directly supervised this project, is now the Dairy Specialist for the province of Manitoba.

ARDI Home - Contact ARDI