Phenotypic associations

The results obtained showed that more than 60 % of animals included in the study have a score equal to or higher than 80 FP, indicating that more than half of the population is in the top two conformation classes (B and VG). Table 2 shows that there is a statistical difference (P<0.001) between the FP classes over MP, PG, PP and NLAC, observing that increasing the FP class increases the production level in three characteristics studied (MP, PG and PP) and the NLAC.

VG animals had 0.26 more lactations than RE animals. This difference represents approximately 79 more days in lactation, which increases the total production of the animals and makes each animal more profitable. These results are reflected in the MP averages of the VG class animals, which had 21% more MP than the RE class. The results in the population under study, coincide with those reported in previous studies conducted in Holstein cattle in Colombia, where they showed that CC are positively correlated with milk production ^{Madrid & Echeverri, 2014}. Unlike productive traits, CC can be early selection tools in dairy cattle, since the qualification can be obtained at the beginning of the first lactation; in addition, it has been reported that these traits present positive genetic correlations with traits of economic interest, such as milk production, milk composition and udder health (^{Duru et al., 2012}).

As for milk, for components (PG and PP), cows with the highest class (VG) had higher production (827 and 729 kg, respectively) although in the case of PP there was no significant difference (P<0.001) between B and VG cows. Results showed that animals in the lowest class (RE) had 22% lower PG and 33% lower PP compared to animals in the highest class (VG). These results corroborate the positive relationship between milk production and components with CC classification (^{Madrid & Echeverri, 2014}) , and corroborate reports of positive genetic correlations (^{Zavadilová & Štt́pková, 2012}) indicating that cows with better anatomical structure, tend to improve their physiological efficiency.

Correlations between conformation systems were moderate to low, highlighting the one between SU and SEC (0.40), between SU and SP (0.29) and, between SEC and SA (0.27) (Table 3).

The magnitude of these correlations shows the anatomical and functional dependence that exists between the systems and the possibility of being improved by selection on an individual basis; since, for example, VPPT and VLPT are related to haunch inclination and udder development and size are highly influenced by SIZ and STA, characteristics of SEC. Correlations between CC systems and productive traits were generally low, with only SP correlations standing out, which presented a correlation of 0.47 with PP and 0.38 with PG. Similar results have not been reported in other populations.

Correlations between genetic values

The correlation found between HTPL and HTPPK was high and positive (0.76), similar to that found with HTPGK (0.69), which means that cows with higher predicted genetic value for milk production tend to possess higher genetic value for total solids production (^{Madrid & Echeverri, 2014}). Correlations found between conformation systems with HTPGK and HTPPK were mostly positive. SEC showed the lowest correlations with HTPL (-0.05), and no significance for HTPGK and HTPPK. The correlation between SU and HTPL was low and positive (0.28), being this the system most correlated with HTPL.

From all the SEC CC, the correlation found between STA and HTPL was the one with the highest value (0.21) (Table 4).

Although the estimates in this study were correlations between genetic values, they were similar to the genetic correlations reported in a population of primiparous Holstein cows in Turkey by ^{Tapki & Ziya GÜZEY (2013)}, which was 0.24 but higher than that reported in another population of the same breed in third calving cows in the same country which was 0.14 (^{Duru et al., 2012}), indicating that the height of animal influences milk production capacity. In barns, it is recommended to select animals with a score of 7 for STA, since these cows have the same capacity to produce milk as larger cows thus reducing the energetic cost of maintenance in these animals. The correlation found between PUNA and HTPL was low and positive (0.08), partially differing with other authors (^{Weller & Ezra, 2016}) who reported negative and low values (-0.04). The correlation between ANCA and HTPL was higher (0.18) and similar to those reported in Turkish Holstein cattle (^{Duru et al., 2012}) where they reported a correlation of 0.19 between the same traits. ANCA determines the separation between the cow's hind feet that the further apart they are, the wider the hind udder will be, allowing it to store and produce a greater amount of milk (^{Getu & Misganaw, 2015}). The correlations between ANCA with HTPGK and HTPPK were low and significantly different p<0.001 (0.09 and 0.08, respectively).

SU had a positive correlation with HTPL (0.28), a result indicating that cows with strong, large and well implanted udders produce higher milk quantity. From SU traits, PU had a low and negative correlation with HTPL (-0.05), (Table 5).

For the same traits, in Holstein cattle from Italy (^{Samoré et al., 2010}) reported negative genetic correlations (-0.37), but lower than those reported in Colombia (^{Corrales et al., 2012}) where they found a negative and high correlation (-0.72). The results show that, in these populations, cows with high milk production potential have a shallower udder depth, since animals with shallow udders have little amount of mammary tissue and therefore their production and storage capacity is limited (^{Piccardi et al., 2012}).

Similar to PU, LONPE presented a low genetic value correlation with HTPL (0.05 respectively; p<0.001). The rest of the SU traits showed a positive and moderate correlation with HTPL, highlighting the one between HTPL with POSPP (0.47) and LM (0.40) (Table 5). These results coincide with those reported in Colombia (^{Corrales et al., 2012}), a population in which a positive correlation between HTPL with AIUP and BF was found, because wide udders with high insertion are related to the amount of mammary tissue and greater milk storage capacity. The relationship between HTPL and POSPP may be largely due to the fact that animals with well positioned teats have greater ease of milking, therefore, they tend to produce more milk and at the same time may have a lower incidence of mammary gland diseases. Correlations between SU characteristics and HTPGK and HTPPK presented similar trends to those of HTPL, results that indicate that good udder conformation contributes to increased milk, fat and protein production capacity (^{Ptak et al., 2011}).

Several authors agree that cows with good locomotion tend to be high milk producers and remain in the herd longer than cows with poor scores for the leg system (^{Wasana et al., 2015}). In this study, correlations of genetic values between SP and HTPGK, and HTPPK were low (0.06 and 0.08 respectively), similar to genetic correlations reported in a Holstein population from Italy (^{Battagin et al., 2013}), who found a correlation of 0.07 with HTPGK and 0.02 with HTPPK. The correlation of SP with HTPL was also low (0.08), different from that reported in the same population, in which they estimated a correlation of 0.24.

The individual SP traits that presented the highest correlation with HTPL were ANPEZ and CALHU (0.22 and 0.34, respectively), while these same CC presented a medium correlation with HTPGK and HTPPK (Table 6).

In Holstein cattle from Spain, ^{Pérez-Cabal & Alenda (2006)}reported a low positive genetic correlation of 0.12 between VPPT and HTPL, results that coincide with those reported in this study, in which a correlation of 0.16 was found. In UK Holstein cattle ^{Ptak et al. (2011)} suggested that feet and hooves are indirectly related to milk production, as cows with poor scores for ANPEZ, VLPT and VPPT show impaired longevity, production and fertility, as well as being traits related to repeat service at 56 days; suggesting that animals with good scores for feet may be less likely to repeat service, reducing days open (^{Wall et al., 2005}). The importance of SP characteristics in production systems lies in the fact that cows have to move daily to be milked and if they do not have good strength and locomotion in the limbs, over time it can cause a deterioration in the productive life (^{Kern et al., 2015}).

The correlation between AG and HTPL found in this study was positive (0.34) (Table 6), similar to that found in the Brazilian Holstein population (^{Campos et al., 2015}) where they reported a correlation (0.38). Similarly, in Turkish Holstein cows, ^{Duru et al. (2012)} reported a positive correlation of 0.21 for the same traits, while in Colombia, ^{Corrales et al. (2012)} reported a lower correlation (0.14). Although the methodology to find the relationship between traits was different, results in Mexico are similar to those presented in Italy in Brown Swiss cattle (^{Samoré et al., 2010}), with a correlation between AG and HTPL of 0.36. With the results found, it can be inferred that more angled cows tend to produce more milk, because animals with greater separation between ribs and greater angle have a better body capacity; in addition, animals tend to have flatter bones, an important aspect in the deposition and extraction of calcium during lactation (^{Carvajal- Hernández et al., 2002}).