Livestock Research for Rural Development 19 (7) 2007 Guide for preparation of papers LRRD News

Citation of this paper

Effect of parity, season and year on reproductive performance and herd life of Friesian cows at Stella private dairy farm, Ethiopia

Gebeyehu Goshu, Kelay Belihu and Abebe Berihun

Addis Ababa University, Faculty of Veterinary Medicine, P.O. Box 34, Debre Ziet, Ethiopia
ggoshu2000@yahoo.com   ;   belihudes@yahoo.com

Abstract

A retrospective study was carried out at Stella Dairy Farm, Ethiopia, using 13 years data (1993 to 2006) to estimate number of services per conception (NSC), days open (DO), calving interval (CI), herd life (HL), number of lactations (NL) and productive age (PA). Fixed factors considered were parity (8+), season (3) and year (13). Data were analysed using the General Linear Model of SAS. Cumulative frequency (%) was used to study the distributions of DO and CI.

The overall mean for NSC, DO and CI were 1.720±0.056 (n=619), 177±5.4 (n=448) days and 456±5.4 (n= 423) days, respectively. Parity had significant effect on NSC (P<0.05) DO and CI (P<0.001). Seasons of conception had no effect on the parameters studied. The overall means (n=105) for HL, NL, and PA were 3048±107days, 3.88±0.23 and 1753±106 days, respectively. Birth year significantly affected the variables. About a third of the heats inseminated required two or more services. About 47% of the DO and 58% of the CI had a length of 130 and 450 days or more, respectively.

To improve the efficiency of reproductive performances and herd life, it is recommended to give attention in heat detection, timely insemination, adequate feeding, proper management of postpartum reproductive problems and early growth management of heifers.

Key words: Ethiopia, fixed factors, Friesian, herd life, reproductive performance, Stella farm


Introduction

Low fertility is of economic importance for dairy enterprises, because it results in higher levels of involuntary replacement and reduced annual milk production. Reports indicated that calving interval of 12 to 13.5 months; number of services per conception of 1. 3 to 1.5 and days open of 85 days are considered as standard values (McDowell 1985; Radostits 2001). Studies following long- term trends in dairy herd reproductive indexes, however, noted a declining performance. A 20-years study of 70 dairy farms in Kentucky showed that the average services per conception performance decreased by 80% but days open increased by 27 days (Silvia 1998). Similarly at Assela farm in Ethiopia the number of services per conception and calving interval increased from 1.92 services and 463 days (Enyew et al 2000) to 2.15 services and 490 days (Yohannes et al 2001), respectively. These necessitate periodic evaluation of the dairy cows' for appropriate corrective measures. The breeding program of some commercial dairy farms including Stella farm was evaluated for number of service per conception and days open (Asseged and Birahnu 2004). However, the effects of parity, season and year were not separated and herd life performances were not determined for individual farm. The objective of this study was to estimate the effects of fixed factors on number of services per conception, days open, calving interval and herd life of dairy cows at Stella farm, Ethiopia.
 

Materials and methods

Study area

Stella farm is located 24km south west of Addis Ababa at an altitude of 2260m. It receives average annual rainfall of 1100 mm. The mean annual minimum and maximum temperatures were 9°C and 25°C, respectively the average being 23.9°C. The average relative humidity is 49.3%. The seasons defined for the area are short rainy season (March to May), long rainy season (June to September) and dry season (October to February).

Management of study animals

Hay conservation was a common practice and was given ad lib. Concentrate feed was prepared fromnoug (Guizotia abyssinica) cake (11%), wheat bran (87.5%) and salt (1.5%). There was routine vaccination against FMD, blackleg, anthrax, pasteurellosis and lumpy skin disease. Treatments for clinical mastitis, tick borne diseases, internal parasites were also given. Heifers reaching 18 months of age and/or attain a body weight of 265 kg were considered to be ready for breeding. Heat detection was the duty of herdsmen and semen of Holstein Friesian bulls is used for artificial insemination. The most important reasons for culling were old age, poor milk production, poor body condition, chronic pneumonia and chronic mastitis. Calves were allowed to get colostrum immediately after birth for two days. During the first month of life calves received 1.5 kg of milk per day. Weaning was at the age of 4 months.

Data collection and analysis

Retrospective data collected from 1993 to 2006 were used to assess number of services per conception (NSC), days open (DO), calving interval (CI), herd life (HL), number of lactation (NL) and productive age (PA). After deleting incomplete record a total of 619, 448 and 423 data were used to estimate NSC, DO and CI, respectively. Herd life, NL and PA were estimated from 105 cows that had disposal records. The HL of a cow was obtained from the difference of disposal and birth dates while HL less age at first calving was taken as PA. Number of lactation was defined as the sum of lactation completed before disposal of the cow. The fixed factors studied were calving parity (8), season (3) and year (13). Preliminary analysis showed that interaction effects of the fixed factors were not significant hence not included in the model. Data were analysed using the following General Linear Model (GLM) procedures of the Statistical Analysis System (SAS 1998).

Yijkl=m+Pj+Sk+Rl+ +eijkl,

Where:

Yijkl= dependent variables (NSC, DO, CI, HL, NL and PA);

m = overall mean;

Pj = the effect of the jth parity of the ith cow;

Sk = effect of the kth season of conception, calving or birth;

Rl= effect of the lth year of birth, conception or calving; and,

eijkl= random error associated with Yijkl observation.

Differences between means were determined when the analysis of variance detected significant effects. Relative and cumulative frequencies were used to summarize the distribution of DO and CI. The average oestrous cycle length of 21 days and a 30-days interval were used as class interval for DO and CI, respectively.
 

Results

The overall means for NSC, DO and CI for the farm were 1.720±0.056, 177±5.4 days, 456±5.4 days, respectively (Table 1).

Table 1. Least squares means for effect of parity on number of services per conception (NSC), days open  (DO) and calving interval (CI).

Parity number

NSC

 

DO (days)

 

CI (days)

n

Means ± SE

n

Means ± SE

n

Means ± SE

Overall

619

1.720±0.056

 

448

177±5.4

 

423

456±5.4

1

166

1.365±0.070b

 

133

212±11a

 

126

486±10.8a

2

135

1.403±0.121b

 

103

194±10ab

 

99

474±11.9a

3

104

1.669±0.159ab

 

75

161±11ab

 

69

431±9.9bc

4

77

1.775±0.176ab

 

56

145±13ab

 

55

426±12.9c

5

56

1.457±0.203b

 

36

156±17ab

 

34

427±18.6c

6

37

1.409±0.213b

 

21

133±21b

 

19

454±23.4bc

7

21

2.058±0.387a

 

12

152±25ab

 

11

439±25.4bc

8+

23

1.820±0.307ab

 

11

137±20b

 

10

434±27.9bc

abc Within variable groups, means followed by the same letter do not differ significantly (P>0.05).

The effect of parity number was highly significant (P<0.001) for DO and CI. Days open and CI decreased as parity number increased until 3. Cows in their first parity demanded 35 and 30 more days than the average for DO and CI, respectively. The effect of parity was significant (P<0.05) for NSC and the same trend was observed in this case too. The first parity cows required 0.36 less services than the population average.

Season did not significantly affect (P>0.05) the reproduction parameters per se (Table 2). Year significantly (P<0.05) affected the NSC, DO and CI. The lowest NSC was recorded in 1993/94 (about 1.16) while the highest was in 2000 (2.2). Higher values of DO and CI were observed in the first year of production.

Table 2.  Least squares means (± se) for effect of season and year on number of services per conception (NSC), days open (DO) and calving interval CI)

Conception season/year

NSC

DO, days

Calving season/year

CI, days

n

Means ± se

n

Means ± se

n

Means ± se

 

 

        NS

 

     NS

 

 

    NS

Short rainy

130

1.533±0.121

96

166±10

Short rainy

122

443±10.4

Main rainy

223

1.587±0.089

166

159±10

Main rainy

160

449±8.9

Dry season

266

1.738±0.087

185

159±8

Dry season

141

447±9.0

 

 

         *

 

     *

 

 

       *

1993

7

1.167±0.264b

5

216±125ab

1993

4

774±137

1994

11

1.154±  0.244b

6

111±31b

1994

8

529±76.2a

1995

15

1.816±0.287ab

9

225±70a

1995

10

402±17.2bc

1996

23

1.488±0.119ab

16

176±42ab

1996

18

440±24.7bc

1997

32

1.433±0.139ab

21

125±12b

1997

18

399±19.5c

1998

42

1.371±0.096ab

26

155±20ab

1998

41

486±19.9ab

1999

65

1.914±0.188ab

45

164±15ab

1999

48

424±12.8bc

2000

68

2.215±0.246a

46

176±20ab

2000

62

462±14.3bc

2001

83

1.587±0.112ab

69

167±13ab

2001

62

448±11.2bc

2002

112

1.960±0.118ab

77

162±11ab

2002

76

432±14.9bc

2004

98

1.887±0.160ab

79

142±13ab

2004

65

435±12.9bc

2005

53

1.898±0.225ab

40

163±14ab

2005

20

452±16bc

2006

10

1.166±0.133b

8

119b±26

-

-

-

abcWithin variable groups, means followed by the same letter do not differ significantly (P>0.05).*=P<0.05; NS= not significant.

Figure 1 shows the distribution of NSC required.

Figure 1.  Distribution of number of services per conception  (n=619) expressed as percentages of all conceptions

Out of the 619 detected heats, 65% conceived in the first insemination and the rest required two or more services per conception. Of the total DO (448) and CI (423) records, only 17 % and 22 % met the standard set for commercial dairy farms, respectively. About 47% of DO and 58% of the CI required 130 (two oestrous cycles) and 450 days or more, respectively (Table 3).

Table 3.   Distribution of days open and calving interval.

Days open (n=448)

Calving interval (n=423)

Interval, days

Relative frequency, %

Cumulative frequency, %

Interval, days

Relative frequency, %

Cumulative frequency,  %

Up to 85

17

17

Up to 365

22

22

86-107

11

28

366-396

14

36

108-129

10

38

397-427

11

47

130-151

9

47

428-458

11

58

152-173

9

56

459-489

12

70

174-195

6

62

490-520

7

77

196-217

8

70

521-550

6

83

218-239

6

76

551-580

5

88

240+

24

100

581+

12

100

The overall means for NL, HL and PA were 3.88±0.23, 3048±107days and 1753±106 days, respectively (Table 4). Birth season had no significant effect on all of the parameters (P>0.05) while year of birth had (P< 0.001).

Table 4.   Least squares means (± se) for herd life (HL), number of lactation (NL) and productive ages (PA).

Source

n

HL, days

NL

PA, days

Means ± se

Means ± se

Means ± se

Overall

105

3048±107

3.88±0.23

1753±106

Birth year

 

***

***

***

1993

15

4771±109a

7.55±0.42a

3385±147a

1994

7

3946±92b

6.11±0.58b

2575±86b

1995

15

3274±83c

4.77±0.37c

2047±91bc

1996

15

3031±110cd

4.43±0.32cd

1832±137c

1997

13

2645±160cde

3.30±0.42de

1335±176cd

1998

11

2432±88def

2.67±0.28ef

997±100d

1999

14

2446±304def

1.61±0.20f

1045±319d

2000

10

1959±315ef

1.64±0.22f

849±310d

2001

5

1959±623ef

1.290±f

886±622d

Birth season

 

NS

NS

NS

Short rainy

31

3210±209

3.97±0.45

1664±203

Main rainy

33

2966±154

3.82±0.38

1666±164

Dry season

41

3210±187

3.85±0.39

1890±185

a..fWithin variable groups, means followed by the same letter do not differ significantly (P>0.05). ***=P<0.001; NS= not significant.

There was a declining trend of the estimates along the studied years. A separate analysis for the 105 cows showed that the age at first calving was 1294 days and about 51.9% of them were culled at early ages
 

Discussion

Number of services per conception

The number of services (1.72±0.056) reported for the present farm is closer to the range given for well managed herd (McDowell 1985) and similar to the estimates for different breed groups at Abernosa Ranch (1.7) (Mekonnen et al 1993), Holeta Research Centre (1.72) (Gifawosen et al 2003) and smallholder dairy farms (1.6) (Lobago et al 2006) but lower than the estimates obtained for crosses in Asela farm (2.15) (Yohannes et al 2001) and commercial diary farms (2.0) (Asseged and Birhanu 2004).

Number of services per conception tended to increase significantly (p<0.05) with parity number (until parity 3). The possible cause of the low NSC for younger cows is not clear and whether that was due to physiological or differential treatment needs to be established. Similar report was made by Nega and Sendrose (2000) at the Holeta state farm that first parity cows had less NSC compared to two and above parity cows. A non significant effect of parity on the NSC were reported by Mekonnen and Goshu (1987) and Asseged and Birhanu (2004).

Season of conception had no significant effect in this study on the NSC. This might be associated to the zero grazing practice in the farm, which makes the effects of seasonal variation in forage developments and feed availability minimal. The absence of seasonal effects on NSC was reported by Mekonnen and Goshu (1987) and Asseged and Birhanu (2004). However, Mekonnen et al (1993) and Gifawosen et al (2003) found a significantly less NSC in the wet season. Year had affected NSC significantly (P< 0.05). The differences observed in NSC between the study years could be related to the variation in feeding and management practices in the years and the same explanations apply to DO and CI. Similar year effect as observed in this study was reported by Mekonnen et al (1993) while Mekonnen and Goshu (1987) and Nega and Sendrose (2000) reported a non-significant effect of calving year.

This study showed that a good proportion of conception cases (about 35%) demanded two or more services (Figure 1).This could suggest the presence of repeat breeders in the herd, occurrences of postpartum reproductive problems, erratic heat detection skills, inefficiency of AI and/or poor body condition (level of feeding) in the herd. Shiferaw et al (2005) found that cows with reproductive disorders required more services per conception and had longer intervals from calving to first service and to conception. Proper and accurate heat detection is a key to efficient reproduction and four to five checks each day to determine the onset of true standing heat gives a better idea when to inseminate. Rocha et al (2001) reported significant differences in non-return rates and calving rates among AI technicians. According to Gebregziabher et al (2005), calving and postpartum body weight gain was inversely related to days to first service. Heavier cows at calving and cows that gained weight during the first three months postpartum were in a positive energy balance, which enabled them to return to normal oestrus cycles. Although not significant, Tadesse and Zelalem (2003) also noted a decrease in the NSC required for cows supplemented high level of protein.

Days open

The overall average DO (177±5.4 days) obtained compares favourably with the 181 and 187 days reported by Yohannes et al (2001) and Asseged and Birhanu (2004), respectively. The DO found in the current study is about twice the limit given to a well managed herd and may be related to the change in management and feeding. The effect of low level of nutrition on extended postpartum period due to weight loss was noted by Gebreegziabher et al (2005). Moreover, Tadesse and Zelalem (2003) reported that increasing the level of protein supplementation from low (2kg/day) to high (4 kg /day) reduced the post partum interval from 159 to 100 days. The highest DO required for the first and second parity cows in this study could be due to the nutritional requirements for growth and their inability to quickly initiate postpartum ovarian activity because of low level of body reserves. Other authors (Mekonnen and Goshu 1987; Yohannes et al 2001 and Gifawosen et al 2003) reported a non significant effect of parity on DO.

Season of conception had not significantly affected the DOprobably for the same reason mentioned above. Significant year effect as in the present case was also reported by several authors (Mekonnen and Goshu 1987; Yohannes et al 2001and Gifawosen et al 2003). The better efficiency noted for NSC was not observed for DO or CI (Table 1). The large deviation of DO from the population mean in the first years might be related to early abortion (because of various reproductive tract diseases) or lack of experience in skills of reproductive and feeding management.

Calving interval

The population mean of 456 days is in agreement with the estimates documented in Ethiopia for F1 crosses ( Teferi 1994), upgraded diary cows (Goshu and Hegde 2003) and Friesian cows (Melaku 1994). The long CI observed could be due to the higher NSC and DO in the farm (Table 3 and Figure 1). Yoseph et al (2000) reported even longer CI in the urban and peri urban dairy production system due to extended lactation length in theses farms to harvest more milk.

First parity cows required longer CI than the third or above parity cows. This may be due to the practice of selective culling against slow breeders in later parities and the additional nutritional requirements of cows in early lactation life for growth. Non significant effect of parity was reported by Yohannes et al (2001). Season of calving had no effect while year of birth of cows had significant effect on the length of CI most likely for similar reasons mentioned above. Similar findings were reported by Yohannes et al (2001) and Goshu and Hegde (2003). About 58% of the records of CI demanded 396 days and 47% proportion of the DO had more than 130 days interval (Table 3) might be because of unnoticed cycles, early loss of embryo and late resumption of ovarian activity. The farm average of 1.72 NSC indicates the possibility of reducing the DO and CI through proper heat detection, feeding and post partum reproduction management.

Herd life, NL and PA

The general goal of dairy farms is for cows to produce a maximum amount of milk per day of life. The overall mean of 3048 days (8.3 years) obtained for HLin this study (Table 4) is higher than the 5.3 years for grade Boran cattle in Tanzania (Trail et al 1985) and the 6.02 years for crossbred cows in Ethiopia (Enyew et al 2000) but in agreement to the 7.9 years for dairy cows in Cheffa farm in Ethiopia (Goshu 2005). The NL obtained in the present farm (3.88) is closer to the findings of Thorpe et al (1994), Enyew et al (2000) and Goshu (2005) who reported about 3.2, 3.2 and 4.2, respectively. Saeed et al (1987) mean number of 4.02 calves for Kenana cattle during their lifetime. Number of calving, age at first calving and the length of productive age are closely related parameters. Nilforooshan and Edriss (2004) noted negative effects of longer ages at first calving on productive life of cows. The age at first calving of 1294 days in the present study indicates that cows calved late in their age, hence had less PA.

Productive age of cows was significantly affected by year of birth of cows in our study, which could be probably due to the variations of management in the years. Faust (1993) reported that more than 90% of differences among cows in productive life are due to management and environment. The progressive decline in NL, HL and PA might be associated to the change in the feeding and health care practices and the high rate of involuntary culling. Study of disposal causes for the same farm showed that 51.9% cows were sold at early age (Abebe 2005) and this could be another reason for low lifetime performance.
 

Conclusions

Acknowledgments

The authors would like to thank the management of Stella farm for allowing us to use the herd data.


References

Abebe Berihun 2005 Herd life and lactation performance of Holstein Friesian cows at Stella dairy farm, Ethiopia. DVM thesis. Faculty of Veterinary Medicine, Debre Zeit, Ethiopia

Asseged B and Birhanu M 2004 Survival analysis of calves and reproductive performance of cows in commercial dairy farms in and around Addis Ababa, Ethiopia. Tropical Animal Health and Production. 36: 663-672.

Enyew Nigussie, Bränäng E and Rottmann O J 2000 Reproductive performance and herd life of crossbred dairy cattle with different levels of European inheritance in Ethiopia. Livestock production and the environment: implication for livelihoods. Proceedings of the 7th annual conference of the Ethiopian Society of Animal Production (ESAP). 26-27 May 1999, Addis Ababa, Ethiopia.

Faust M 1993 Capitalizing on dairy herd life. U. S. National Dairy Data base

Gebeyehu Goshu 2005 Breeding efficiency, lifetime lactation and calving performance of Friesian-Boran crossbred cows at Cheffa farm, Ethiopia. Livestock Research for Rural Development. Volume17, Article#7. http://www.cipav.org.co/lrrd//lrrd17/7/gosh17073.htm

Gebeyehu GoshuG and Hegde B P 2003 Age at first calving, calving interval and milk yield performance of Friesian-Boran crossbred cattle at Cheffa state farm, Wollo, Ethiopia. Bulletin of Health and Production in Africa. 51:190-197.

Gebregziabher Gebreyohannes, Azage Tegegne, Diedhion M L and Hegde B P 2005 Days to first service, conception rate and service period of indigenous and crossbred cows in relation to postpartum body weight change at Bako, Ethiopia. Ethiopian Journal of Animal production. 5(1): 83-90.

Gifawosen Tessema, Alemu Geberewold, Azage Tegegne, Diediou M L and Hegde B P 2003 Study on reproductive efficiency of Boran and its crosses at Holetta research farm: Effect of genotype, management and environment. Ethiopian Journal of Animal Production. 3(1) 89-108.

Lobago F, Bekana M and Gustafsson H 2006 Reproductive performances of dairy cows in smallholder production system in Selalle, central Ethiopia. Tropical Animal Health and Production. 38: 333-342.

McDowell R E 1985 Crossbreeding in tropical areas with emphasis on milk, health and fitness. Journal of Dairy Science. 68: 2418-2435.

Mekonnen Hailemariam and Goshu Mekonnen 1987 Reproductive performance of Fogera cattle and their Friesian crosses. Ethiopian Journal of Agricultural Science. 9(2): 95-114.

Mekonnen Hailemariam, Keno Banjaw, Tefera Gebremeskel and Hiskias Ketema 1993 Productivity of Boran cattle and their Friesian crosses at Abernossa ranch, rift valley of Ethiopia. I. Reproductive performance and preweaning mortality. Tropical Animal Health and Production. 25: 239-249

Melaku Negash 1994 Reproductive performance of a Holstein-Friesian dairy cattle herd at Holeta, Shoa, Ethiopia. M.Sc thesis. Alemaya University of Agriculture, Dire Dawa, Ethiopia.

Nega Tolla and Sendrose Demeke 2000 Effect of calving year, season, age and parity classes on production and reproduction performance of Holstein-Friesian cows at Holetta state farm. Pastoralism and agro pastoralism: which way forward. Proceedings of the 8th annual conference of the Ethiopian Society of Animal Production (ESAP). 24-25 August 2000, Addis Ababa, Ethiopia.

Nilforooshan M A and Edriss M A 2004 Effect of age at first calving on some producitve and longevity traits in Iranian Holsteins of the Isfahan province. Journal of Dairy Science. 87(7): 2130-2135. http://jds.fass.org/cgi/reprint/87/7/2130

Radostits O M 2001 Herd Health: Food Animal Production Medicine. Third Edition. Philadelphia. pp 255-289.

Rocha A, Rocha S and Carvalheire J 2001 Reproductive parameters and efficiency of inseminators in dairy farms in Portugal. Reproduction of Domestic Animal. 36(6): 319-324.

Saeed A M, Ward P N, Light D, Durkin J W and Wilson R T 1987 Characterization of Kenana cattle at Um Banein, Sudan. ILCA Research Report 16. International Livestock Center for Africa (ILCA), Addis Ababa, Ethiopia.

SAS (Statistical Analysis System) 1998 SAS Institute Inc., Cary, NC, USA.

Shiferaw Y, Merga B and Tesfu K 2005 Reproductive disorders of crossbred dairy cows in the central highlands of Ethiopia and their effect on reproductive performance. Tropical Animal Health and Production. 37(5): 427-441.

Silvia W J 1998 Changes in reproductive performance of Holstein dairy cows in Kentucky from 1972 to 1996. Journal of Animal Science 76 (Supplement 1): 244.

Tadesse Bekele and Zelalem Yilma 2003 Feeding noug 'Guizotia abyssinica' cake as protein source to lactating Borana ´ Jersey crossbred cows: performances in milk yield, reproduction and feed efficiency. Farm animal biodiversity: status and prospects. Proceedings of the 11th annual conference of the Ethiopian Society of Animal Production (ESAP). 28-30 August 2000, Addis Ababa, Ethiopia.

Teferi Dina 1994 Reproductive and milk yield performance of crossbred dairy cattle at the cooperative farms in the Arsi region, Ethiopia. M. Sc. thesis. Alemaya University of Agriculture, Dire Dawa, Ethiopia.

Thorpe W, Morris C A and Kang'Ethe P 1994 Crossbreeding of Ayrshire, Brown Swiss, and Sahiwal cattle for annual and lifetime milk yield in the lowland tropics of Kenya. Journal of Dairy Science. 77: 2415-2427. http://jds.fass.org/cgi/reprint/77/8/2415?ck=nck

Trail J C M, Sones K, Jibbo J M C, Durkins J, Light D E and Murry M 1985 Productivity of Borana cattle maintained by chemoprophylaxix under trypanosomosis risk. ILCA Research Report 9. International Livestock Center for Africa (ILCA), Addis Ababa, Ethiopia

Yohannes Afework, Azage Tegegne and Tesfu Kassa 2001 Reproductive performance of crossbred dairy cows at Asella Livestock Research Station, Arsi, Ethiopia. Ethiopian Journal of Animal Production. 1: 1-12.

Yoseph Mekasha, Azage Tegegne, Alemu Yami and Umnna N 2000 Reproduction management and reproductive performance of dairy herds in urban and peri urban dairy production systems in Addis Ababa milk shed. Pastoralism and agro pastoralism: which way forward. Proceedings of the 8th annual Conference of the Ethiopian Society of Animal Production (ESAP). 24-25 August 2000, Addis Ababa, Ethiopia.



Received 19 December 2006; Accepted 22 May 2007; Published 6 July 2007

Go to top