Udder and teat morphology traits associated with milk production and somatic cell score in dairy sheep from a New Zealand flock

ABSTRACT

The objective of this study was to estimate the phenotypic correlations between udder and teat morphology traits, milk production traits, and somatic cell score in dairy sheep from a flock in New Zealand. A total of 162 lactating ewes were scored for morphology traits during the milk production season of 2021–2022. The 130-d lactation yields of milk, fat, protein, and lactose were obtained with 2–4 test-days from each ewe and modelled using random regression with orthogonal polynomials. Age had a significant effect on all udder and teat traits. Coat-colour (genetic variety within the breed; white or black) was a significant effect for teat angle and udder separation. Udders that were above the hook were associated with lower milk, fat, protein, and lactose yields. Udders with well-defined separation between halves were associated with higher milk, protein, and lactose yields, and with lower somatic cell count. Well-attached udders were associated with lower fat yield and lower somatic cell score. Teats with a backwards angle were associated with lower milk and lactose yields. Further studies are needed to estimate heritability and genetic correlations between these traits to determine whether these traits should be implemented in breeding programs for dairy sheep in New Zealand.

KEYWORDS:

• Dairy sheep
• udder morphology traits
• teat morphology traits
• milk production
• somatic cell score

Introduction

The New Zealand dairy sheep industry is relatively new. This sector started in 1992 with the importation of a small pool of East Friesian genetics (Lees and Lees 2016) that were crossed with other breeds. More recently, Awassii and Lacaune have also been incorporated in some New Zealand flocks (Peterson and Prichard 2015; Ministry of Primary Industries – Massey University 2020). Although small, this industry is growing due to the high nutritional value of sheep milk seen by the higher concentrations of proteins, fats, vitamins, and minerals when compared to cow and goat milks (Park et al. 2007), and due to the potential to be used as functional food and promote human health (Mohapatra et al. 2019). There is also potentially lower environmental impact caused by dairy sheep farming when compared to the traditional dairy cow farming (Downie-Melrose 2014; Smith et al. 2018).

In New Zealand, as in many other countries, dairy animals (including sheep, cows, goats, deer) are milked in parlours with machinery that uses mechanical pulsation and clusters (cups) that attach to the teats. Poor mammary morphology is mainly characterised by long pendulous udders and by teats positioned too laterally or too vertically, which is not desirable for cluster attachment during milking or for suckling lambs, respectively. Therefore, mammary morphology is crucial at the farm and can affect the everyday efficiency of milking and of nursing of lambs. Also, udder and teat traits can have moderate to high heritability, depending on the breed, and these traits have been included in the selection index of some dairy sheep populations around the world (Fernández et al. 1997; Legarra and Ugarte 2005; Casu et al. 2006).

Furthermore, it has been reported that selection of dairy sheep solely for milk yield can indirectly result in impaired udder morphology for machine milking and lamb suckling (Legarra and Ugarte 2005; Marie-Etancelin et al. 2005), reduced milk solids (Bencini 2002), increased somatic cell count (Fernández et al. 1997; Legarra and Ugarte 2005), and reduced fertility (David et al. 2008).

There is limited information available in the scientific literature about genetic programmes for dairy sheep in New Zealand. One study using in a single flock has proposed a model for genetic evaluation of dairy sheep in New Zealand (Scholtens et al. 2018), without considering udder and teat morphology traits. However, most farms still rely on their own selection schemes. More recently, large New Zealand dairy sheep companies have started running their own genetic programmes claiming to have a breeding goal and a selection index that includes udder conformation traits (New Zealand Sheepbreeders Association 2023a, 2023b) alongside milk yield, and milk composition traits. Nationally, there is still not one defined breeding goal for dairy sheep.

Only a few studies have investigated udder traits in New Zealand sheep, and these were done on non-dairy sheep, where udder depth was phenotypically associated with lamb survival to weaning (Griffiths et al. 2019), and lamb weight and milk volume (Yusuf et al. 2018). The aim of this study was to investigate the phenotypic correlations between udder and teat morphology traits, milk production traits and somatic cell score (SCS) in dairy sheep from a New Zealand flock. Knowledge of the relationship between udder morphology traits and milk production is valuable to inform the New Zealand dairy sheep industry for further genetic studies on these traits, using larger populations of dairy sheep.

Materials and methods

Test-day records were gathered during the milk production season of 2021-2022, from 162 ewes from a commercial farm, in Masterton, New Zealand. An animal ethics approval was obtained for this study (Massey University Animal Ethics Committee-Protocol 21/45). The breed used in the current study is mainly formed by East-Friesian, with some Coopworth and Border Leicester bloodlines. In this flock, ewes can distinctly be separated by coat colour, black and white.

The flock was milked twice-a-day (TAD) in October. In November, milking frequency was shifted to once-a-day (OAD) to align with the seasonal availability of pasture, this is a common practice on this farm. Therefore, ewes that had their lambs weaned after 31 Oct 2021 were only milked once-a-day. To adjust for the milking frequency, days in TAD milking were calculated for each ewe (1st of November−weaning date).

Milk yield was recorded on the test-day, and a representative milk sample was taken for composition analyses, as described in Marshall et al. (2023). The lactation curves were modelled using 2–4 test-day records from each ewe. The 130-day lactation yields of milk, fat, protein, and lactose were obtained using random regression with orthogonal polynomials (Marshall et al. 2023). Time was defined as d = days in milk−35, as most records were made 35 days after the lambing date due to the exclusive suckling period. Legendre polynomials were used to standardise values to the interval [−1, … ., 1], and the coefficients were then calculated using the Rodrigues formula (Askey 2005): $P0(t)=1,$ $P1(t)=x,$ $P2(t)={\displaystyle \frac{1}{2}(3x^2-1),}$ $P3(t)={\displaystyle \frac{1}{2}(5x^3-3x),}$ $P4(t)={\displaystyle \frac{1}{8}(35x^4-30x^2+3),}$ $P5(t)={\displaystyle \frac{1}{8}(63x^5-70x^3+15x)}$

where x = −1 + 2, and ${\displaystyle \frac{(t-t_{min})}{(t_{max}-t_{min})}}$, with t_min = 1 and t_max = 130.

Day 1 corresponded to day 35 of lactation.

The random regression model was represented as follows: $y_{\mathrm{ti}}=({\beta }_0{P}_0+{\beta }_1{P}_1+{\beta }_2{P}_2+\dots +{\beta }_n{P}_n)+({\alpha }_{0i}{P}_0+{\alpha }_{1i}{P}_1+{\alpha }_{2i}{P}_2+\dots +{\alpha }_{\mathrm{ni}}{P}_n)+e_{\mathrm{ti}},$where y_ti is the observed yield of animal i at day t, $\beta$’s are the fixed regression coefficients of the lactation curve of the population, $\alpha$ values are random regression coefficients describing the lactation curve for animal i, n is the maximum polynomial order, and e_ti is the random residual of observation y_ti. The estimates of $\beta$ and $\alpha$ were obtained using the MIXED procedure of SAS version 9.4 (SAS Institute Inc., Cary, NC, USA, SAS 2004) with the COVTEST option for covariance parameter estimates. Based on the Akaike (AIC) and Bayesian (BIC) information criteria (smallest is the best), an orthogonal polynomial of order 4 was considered the best fit for modelling daily milk and lactose yields. An orthogonal polynomial of order 5 was considered the best fit for modelling fat and protein daily yields.

The daily milk yield was predicted from 35 to 164 days in milk (or from t = 1 to t = 130) and were then summed to obtain an estimated total milk yield produced by each ewe for 130 days (from day 35 to day 164 of lactation).

Somatic cell score for each ewe was averaged from the somatic cell score obtained from each test-day, calculated as SCS = Log₂(somatic cell count). Throughout the current study, no ewes presented any clinical signs of udder or teat inflammation.

Udder and teat morphology were visually assessed on the 162 lactating ewes once in the season, in December 2021. At the time of assessment, the ewes were in mid to late lactation stages and were being milked once-a-day, in the afternoons. Udder and teat assessments were done immediately prior to milk collection. Selection on teat and udder morphology traits and on milk production traits has been in place for over 7 years in this farm. High repeatability for teat and udder morphology traits within a season (or lactation) has been reported for other breeds around the world, such as Sarda (Casu et al. 2006), Lacaune (Marie-Etancelin et al. 2003), and Churra (de la Fuente et al. 1996), meaning that a single score in the season should be reliable. Assessment of most teat and udder morphology traits was performed from behind the animals as they stood in their bays in the milking shed, except for teat angle, which was performed laterally. These traits were linear scores from 1 to 9 (Figure 1) and include teat position (TP), teat angle (TA), udder depth (UD), udder attachment (UA), and udder separation (US). The desirable value for some udder traits is, in some cases, the highest score or the intermediate score. For udder depth, for example, given its positive relationship with milk production an intermediate score is preferable (Caja et al. 2000).

Figure 1. Linear scoring system of udder and teat morphology traits used in ewes from a commercial flock during the production season of 2021–2022. Teat position = the distance between the teats and the lowest part of the udder (1 = horizontal; 9 = vertical). Teat angle = looking from the side of the animal (1 = pointing forward; 9 = pointing towards the rear). Udder depth = the distance between the udder cleft and the abdominal wall, taking as reference point the line joining the hocks (1 = deep; 5 = hock level; 9 = shallow). Udder attachment = the ratio between udder height and udder width (1 = width smaller than height; 7 = width equals height, or square; 9 = width larger than height). Udder cleft or udder separation = degree of separation between left and right mammary glands (1 = separation is missing; 5 = average separation; 9 = well marked separation).

As there is no single international scoring system in place for sheep udder and teat traits, this study has defined them based on a combination of guidelines from the International Committee for Animal Recording (ICAR 2018) for the different breeds of dairy sheep, and from previous publications for dairy goats (McLaren et al. 2016) and dairy sheep (Casu et al. 2006; Griffiths et al. 2019), whilst aiming similarity with the scoring system that is already in place for dairy cows in New Zealand (Advisory Committee on Traits Other than Production 2020).

Descriptive statistics (mean, standard deviation, and coefficient of variation) for traits were obtained in SAS version 9.4 software (SAS Institute Inc., Cary, NC, USA, SAS 2004). Analysis of variances for the dependent variables udder and teat morphology traits, milk production traits (milk yield, fat yield, protein yield, lactose yield), and SCS were performed in SAS using the MIXED procedure with a linear model that included the fixed effects of ewe coat colour as an indicator of genetic variety within the breed (categorical variable with two levels: black or white ewe), litter size (categorical variable with two levels: 1 lamb or 2 lambs and greater), ewe age (categorical variable with four levels: 1, 2, 3, and 4 years and older), and days in twice-a-day (TAD) milking as covariate, and random residual error. Least squares means and standard errors for each class of the fixed effects were obtained and used for mean comparisons using Fisher’s least significant difference.

Using the solutions of the mixed model, best linear prediction functions were used to predict values of dependent traits at different days milked TAD after weaning.

Partial correlation coefficients between traits were obtained through multiple analysis of variance, using the option MANOVA of the GLM procedure with the same linear model described above. These partial correlation coefficients can be considered as phenotypic correlations because these coefficients are adjusted by fixed effects included in the linear model.

Results

Average milk production per ewe estimated for 130 days of milking, and average udder and teats scores are presented on Table 1. Coefficients of variation for milk production traits were high and ranged from 27 to 31%, and for teat and udder morphology traits ranged from 11 to 25%, being the lowest variation for udder depth (11%) and teat angle (13%).

Table 1. Descriptive statistics^a and F-values for factors affecting milk production (estimated lactation yields of milk, fat, protein, and lactose) and udder and teat morphology traits in dairy sheep from a New Zealand flock during the milk production season of 2021–2022 (N = 162).

Trait	Mean	SD	Min	Max	CV (%)	F-value
						Coat colour	Age	Litter size	Days in TAD^c
Lactation yields^b (kg)
Milk	86.1	26.1	25.1	168.8	30	0.07	6.70 ***	0.63	19.86 ***
Fat	5.1	1.4	2.0	9.5	28	0.61	3.83 **	0.00	27.47 ***
Protein	4.5	1.2	1.7	8.2	27	0.01	4.38 **	1.12	20.51 ***
Lactose	4.1	1.3	1.2	8.0	31	0.03	6.71 ***	0.63	17.48 ***
Average SCS	16.9	1.7	14.1	23.0	10	0.94	1.70	1.48	0.79
Teat morphology
Teat angle	5.5	0.7	4	7	13	12.22***	6.82***	2.76	0.09
Teat position	5.6	1.4	1	9	25	2.98	6.47***	0.03	2.87
Udder morphology
Udder depth	7.6	0.8	5	9	11	0.45	10.15***	0.90	0.47
Udder attachment	7.1	1.6	2	9	23	0.00	22.89***	0.88	1.83
Udder separation	6.7	1.6	1	9	24	5.26**	6.96***	0.10	0.73

^a SD = standard deviation, Min = minimum value, Max = maximum value, CV = coefficient of variation.

^b Lactation yields estimated from daily yields from 35 to 164 days of lactation; Average SCS = average of somatic cell scores (SCS) during the lactation where SCS = Log₂(somatic cell count); Teat and udder morphology traits were scored from 1 to 9.

^c Days in twice-a-day milking.

Statistical significance is given as: * p < 0.05; ** p < 0.01; *** p < 0.001.

The F- and p-values for fixed effects are also presented in Table 1. Age and days in TAD milking significantly affected yields of milk, fat, protein, and lactose. The factor that had the greatest effect on the total yields was days in TAD milking (largest F-values). Litter size and coat-colour did not have any significant effect on milk production. SCS was not affected by any of the fixed effects.

All teat and udder morphological traits were strongly affected by age (p < 0.001), especially udder attachment and udder depth, with F values of 22.89 and 10.15, respectively. Coat-colour significantly affected teat angle (p < 0.001) and udder separation (p < 0.01). Litter size and days in TAD milking did not significantly affect any of the teat and udder morphology traits (Table 1).

Least square means for total lactation yields of milk, fat, protein, and lactose and least square means for teat and udder morphology traits are presented in Table 2. One-year-old ewes produced significantly less (p < 0.05) milk, fat, protein, and lactose than older ewes (Table 2). Three-year-old ewes produced the highest yields of milk, protein, and lactose, whereas four-year-old ewes produced the highest yield of fat. Ewes that lambed late in relation to the median lambing date of the flock missed TAD milking in early lactation (days in TAD = 0) and produced significantly less (p < 0.05) milk, fat, protein, and lactose yield than early lambing ewes that were milked TAD in early lactation (days in TAD ≥ 14). Four-year-old ewes and older ewes had significantly lower (p < 0.05) scores for udder attachment and udder separation.

Table 2. Least-squares means ± standard errors of milk production traits (estimated lactation yields of milk, fat, protein, and lactose) and udder and teat morphology scores for different ewe ages (year), litter sizes, coat colour, and days in twice-a-day (TAD) milking, in dairy sheep from a New Zealand flock during the milk production season of 2021–2022.

Effect	N	Milk	Fat	Protein	Lactose	TA	TP	UD	UA	US
Age
1	23	64.3^b ± 5.4	4.1^b ± 0.30	3.6^b ± 0.30	3.1^b ± 0.30	5.86^a ± 0.54	6.30^a ± 0.97	8.26^a ± 0.61	8.21^a ± 0.59	7.26^a ± 1.35
2	46	85.8^a ± 3.6	5.0^a ± 0.20	4.5^a ± 0.17	4.2^a ± 0.18	5.84^a ± 0.55	5.97^a ± 1.12	7.89^a ± 0.67	7.67^a ± 0.81	7.02^a ± 1.30
3	38	92.1^a± 3.6	5.2^a ± 0.20	4.7^a ± 0.17	4.4^a ± 0.18	5.44^b ± 0.55	5.52^b ± 1.13	7.65^b ± 0.74	7.50^a ± 1.10	6.97^a ± 1.30
≥ 4	55	90.2^a± 3.0	5.3^a ± 0.17	4.6^a ± 0.14	4.3^a ± 0.15	5.21^b ± 0.83	5.10^b ± 1.79	7.16^c ± 0.87	5.78^b ± 1.96	5.98^b ± 1.85
Litter size
1	64	81.7± 2.8	4.9 ± 0.16	4.3 ± 0.14	3.9 ± 0.14	5.48 ± 0.77	5.81 ± 1.43	7.82 ± 0.82	7.43 ± 1.45	6.92 ± 1.51
≥2	98	84.5± 2.9	4.9 ± 0.16	4.4 ± 0.14	4.1 ± 0.14	5.58 ± 0.67	5.50 ± 1.42	7.52 ± 0.84	6.82 ± 1.72	6.54 ± 1.63
Coat colour
Black	32	82.5± 4.0	4.8 ± 0.22	4.3 ± 0.20	3.9 ± 0.20	5.03^b ± 0.82	5.71 ± 1.32	7.40 ± 0.97	6.68 ± 1.57	7.03^a ± 1.33
White	130	83.7± 1.9	5.0 ± 0.11	4.4 ± 0.09	4.0 ± 0.10	5.66^a ± 0.62	5.60 ± 1.46	7.70 ± 0.80	7.16 ± 1.66	6.60^b ± 1.64
Days in TAD
0	53	71.2^b± 3.2	4.1^b ± 0.18	3.8^c ± 0.15	3.4^b ± 0.16	5.60 ± 0.71	5.50 ± 1.46	7.69 ± 1.04	7.07 ± 1.94	6.67 ± 1.64
14	36	86.8^a± 4.0	5.1^a ± 0.22	4.5^b ± 0.19	4.2^a ± 0.20	5.66 ± 0.67	6.02 ± 1.44	7.66 ± 0.75	7.16 ± 1.27	7.00 ± 1.47
21	37	86.3^a± 4.0	5.3^a ± 0.22	4.4^b ± 0.19	4.2^a ± 0.20	5.32 ± 0.78	5.59 ± 1.40	7.75 ± 0.72	7.40 ± 1.16	6.67 ± 1.52
28	36	94.8^a± 4.2	5.6^a ± 0.24	5.0^a ± 0.20	4.5^a ± 0.21	5.55 ± 0.65	5.41 ± 1.38	7.41 ± 0.69	6.61 ± 1.87	6.41 ± 1.69

N = number of ewes within each category. Days in TAD = 0 (ewes that were only milked once-a-day). Days in TAD = 14 (ewes that were milked for 14 days in twice-a-day milking). Days in TAD = 21 (ewes that were milked for 21 days in twice-a-day milking). Days in TAD = 28 (ewes that were milked for 28 days in twice-a-day milking). ^a,b,c Means with different superscripts within effect are significantly different (p < 0.05).

Phenotypic correlations $(r_P)$ between milk production, SCS, and udder and teat morphology traits are presented on Table 3. Correlations among milk production traits (milk, fat, protein, and lactose yields) were significant (p < 0.001), positive, and high ($r_P$>0.9). Correlations among all teat and udder traits were also positive, these correlations were significant (p < 0.001), and moderate between most teat and udder morphology traits, ranging from 0.31 to 0.61 (Table 3).

Table 3. Phenotypic correlations between estimated lactation yields (milk, fat, protein, and lactose), SCS, and udder and teat morphology traits in dairy sheep from a New Zealand flock during the milk production season of 2021–2022.

Trait^a	Milk	Fat	Protein	Lactose	SCS	Teat angle	Teat position	Udder depth	Udder attachment
Fat	0.92***
Protein	0.97***	0.92***
Lactose	0.99***	0.91***	0.96***
SCS	−0.09	−0.13	−0.05	−0.13
Teat angle	−0.16*	−0.15	−0.14	−0.17*	−0.04
Teat position	−0.05	−0.09	−0.02	−0.05	0.01	0.35***
Udder depth	−0.35***	−0.36***	−0.35***	−0.34***	−0.09	0.36***	0.12
Udder attachment	−0.12	−0.17*	−0.12	−0.12	−0.24**	0.31***	0.32***	0.61***
Udder separation	0.19*	0.12	0.19*	0.20*	−0.18*	0.05	0.37***	0.14	0.34***

^a Milk, fat, protein, and lactose are lactation yields estimated from daily yields from 35 to 164 days of lactation; SCS is the average of somatic cell scores (SCS) during the lactation where SCS = Log2(somatic cell count); Teat and udder morphology traits were scored from 1 to 9.

Statistical significance is given as: * p < 0.05; ** p < 0.01; *** p < 0.001.

All teat and udder morphology traits, except for udder separation, had negative correlations with milk production traits (milk, fat, protein and lactose yields) (Table 3). The correlations between udder depth and milk production traits were all significant (p < 0.001) and moderate, ranging from −0.34 to −0.36. Other correlations between udder/teat morphology and milk production traits were weak and/or non-significant, ranging from −0.17 to 0.20.

Somatic cell score had a negative correlation with teat angle, udder depth, udder attachment, and udder separation. The correlation between SCS and udder attachment was significant (p < 0.01) and moderate (−0.24), and the correlation with udder separation was also significant (p < 0.05), but low (−0.18).

Discussion

Udder depth was the main morphological trait correlated with all milk production traits, with significant (p < 0.001) moderate phenotypic correlations with yields of milk, fat, protein, and lactose ($r_P$ ranging from −0.34 to −0.36). Higher udders were related to lower milk production, with age included as one of the fixed effects. This study agrees with the negative phenotypic relationship between udder depth and milk production reported in various breeds of dairy sheep (Casu et al. 2000; Kominakis et al. 2009; Pourlis 2020; Vrdoljak et al. 2020). In this study, udder depth scores only ranged from 5 to 9, meaning that no ewe had very low udders (below the hock line). This could be a consequence of selection in this flock for udder depth. Intermediate scores for udder depth would be desirable for both lamb suckling and machine milking (cluster attachment) without compromising milk productivity. All other teat and udder morphological traits only had low phenotypic correlations ($r_P$≤0.2) with the milk production traits.

High level of udder attachment for machine milking was associated with lower SCS. High level of udder attachment reflects the better support of the udder given by the lateral suspensory ligaments. Udders that are not pendulous (with less movement) and that are far off the ground, are less prone to damage (bruising and laceration) and are less prone to contamination in the teat canal, preventing infection of the mammary glands, reducing the occurrence of subclinical mastitis. Udder depth and udder attachment were highly correlated traits ($r_P$ = 0.61), so it would be preferable to use udder attachment as it is a more complete type of assessment that includes width and height. Also, variation for udder attachment was higher than variation for udder depth in this flock, which could be translated into more opportunity for selection on udder attachment, but further studies are needed to confirm animal genetic variance for these traits.

Udder attachment also had a low negative correlation fat yield ($r_P$ = −0.17), agreeing with the findings of McKusick et al. (1999) for East Friesian crossbred ewes, where for each cm increase in cistern height there was a relative increase of 0.12% units of milk fat. Ewes with deeper cisterns can store milk and milk fat in the cistern between milkings and avoid the deleterious effects of residual milk on the secretory alveoli of the udder (McKusick et al. 1999).

Udders with good separation between halves (groove between mammary glands) also had a weak but significant (p < 0.05) negative correlation with SCS, agreeing with others (Fernández et al. 1997; Legarra and Ugarte 2005). Good udder separation is a reflect of better support given by the medial suspensory ligament (Casu et al. 2006). Udders with well-defined separation had a low positive correlation ($r_P$≤0.20) with the milk production traits. These correlations were significant (p < 0.05) with milk, protein, and lactose yields.

All udder and teat morphology traits were positively correlated amongst themselves agreeing with the phenotypic correlation results reported by others (Fernández et al. 1995; Kominakis et al. 2009). Further studies are needed to investigate the genetic correlations between udder and teat morphology traits to confirm whether improvement in one of the teat and udder morphological traits would also result in improvement of the other traits. Also, a principal component analysis would indicate the group of teat and udder morphological traits that largely explain variance in the dataset.

Age strongly affected all udder and teat conformation traits, especially udder attachment and udder depth, given the high F values (Table 1), agreeing with Serrano et al. (2002). Low udders, of poor attachment, with poor separation, with lateral teats, and with teats pointing forward were more commonly observed after the third parity (Table 2). High scores for udder and teat conformation traits were more commonly observed on ewes of first and second parities. This is expected, as worsening of the teat and udder conformation traits with the parities has been reported (Casu et al. 2006). To confirm repeatability of overall udder conformation with ageing, principal component analysis across seasons would be useful.

This study confirms that teat and udder morphology scores have negative phenotypic association with milk production. However, intermediate scores for teat and udder conformation traits were good enough for machine milking and for acceptable levels of SCS, without compromising milk production. It is recommended that dairy sheep farmers implement a recording system of udder and teat conformation scores, as well as accurate recording of test-day milk yield and milk composition for the purposes of genetic selection. In addition to this, the New Zealand dairy sheep industry should consider a central database for performance recording of dairy sheep. This would enable nation-wide genetic evaluation on these traits and provide genetic information for commercial farmers.

Conclusion

This was the first study to describe teat and udder traits in dairy sheep in a commercial flock in New Zealand. The results from the multivariate analysis indicate that udder and teat conformations that are highly desirable for machine milking are correlated with lower milk production. Therefore, the selection of animals on machine milking traits should be carefully planned, so that milk yield is not compromised. However, further studies are needed to define heritability for teat and udder morphology traits, and genetic correlations between teat and udder morphology traits with milk production traits, for possible implementation in selection schemes for dairy sheep in New Zealand.

Acknowledgements

This work was also supported by the Riddet Institute National Centre of Research Excellence through a PhD scholarship to A.C.M. The authors would like to thank the commercial farm and the scientific communities from the NZ3M group, Massey University, and AgResearch.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by the New Zealand Ministry of Business, Innovation and Employment (MBIE), through the New Zealand Milks Mean More (NZ3M) Endeavour Programme (Contract MAUX1803).